How Oncologists’ Communication Style influences (Analogue) Patients’
Psychophysiological Response and their Recall of Information: An
Experimental Study
Masterthesis: N. M. (Niki) Medendorp July, 2015
10875654
Academic Medical Center Department of Medical Psychology
First assessors: N. C. (Leonie) Visser, MSc and M. A. (Marij) Hillen, PhD
&
University of Amsterdam
Faculty of Social and Behavioral Sciences
Abstract
During oncological consultations, patients often show high levels of arousal and strong emotions, which might deteriorate the recall of provided information and complicates decision making. A communication style that evokes trust by enhanced conveyance of competence, honesty and caring could reduce patients’ stress. In the present study we
investigated whether using a high trust-evoking communication style by oncologists decreases patients’ psychophysiological response and increases recall of information. Dutch students (N = 97), acting as analogue patients, were randomly assigned to two conditions to watch a scripted video-taped oncological outpatient consultation in which the oncologist employed either a standard or high trust-evoking communication style. Analogue patients’ physiological response, i.e. electrodermal activity (both skin conductance level and responses) and heart rate, were measured. Afterwards, their active and passive recall of information were assessed. The heart rate response was significantly smaller (p = .037) in the high trust-evoking
condition (M = 70.4) than in the standard condition (M = 71.7). Electrodermal activity did not differ between the two conditions for both skin conductance level (p = .989) and skin
conductance responses (p = .512). Active recall of information was significantly higher (p = .039) in the high trust-evoking condition (65.3% correctly answered versus 59.5% in the standard condition), but recognition (passive recall) did not differ significantly (p = .502). In conclusion, a high trust-evoking communication style can decrease patients’ heart rate and increase patients’ active recall. Future research should further clarify the relationship between communication style, psychophysiological response and recall of information .
Keywords: oncology; trust; video vignettes; analogue patients; skin conductance; heart
How Oncologists’ Communication Style influences (Analogue) Patients’ Psychophysiological Response and their Recall of Information: An Experimental Study
Receiving a cancer diagnosis has a major impact on patients and their relatives (Sep, Van Osch, Van Vliet, Smets, & Bensing, 2014). It creates uncertainty because of the modification of the perspective of the patient’s future and the need to make important decisions (Gabrijel et al., 2008; Jansen et al., 2010; Sep et al., 2014). During oncological consultations patients often show high levels of arousal and strong emotions due to mental stress (Sep et al., 2014; Van Osch, Sep, Van Vliet, Van Dulmen, & Bensing, 2014). Mental stress activates the sympathetic nervous system (SNS) (Kreibig, 2010). Activation of the SNS induces the fight-flight response, which increases physiological arousal and prepares the body for action (Schwabe, Joëls, Roozendaal, Wolf, & Oitzl, 2012). Subsequently, stress-induced arousal could influence patients’ memory of provided information according to the inverted U-shape function of stress and the attention narrowing hypothesis (Easterbrook, 1959; Salehi, Cordero, & Sandi, 2010). The inverted U-shape function of stress indicates that cognitive performance is best when an individual is under optimal stress; performance would be impaired under conditions above or below optimal stress levels (Salehi et al., 2010). High stress levels during oncological consultations are likely to exceed the optimum and could therefore lead to limited recall of information. In addition, the attention narrowing hypothesis suggests that a negative or threatening stimulus causes emotional arousal, which attracts attention to the arousing stimulus (Easterbrook, 1959; Guillet & Arndt, 2009). Because the capacity of attention is limited, the hyper attention given to the arousing stimulus reduces the amount of attentional resources available to process other information present in the same context (Guillet & Arndt, 2009; Van Osch et al., 2014). Therefore, the attention-narrowing hypothesis predicts that attention in oncological consultations is directed to the distressing facts particular for that patient, resulting in a decline of recall about other provided information (Easterbrook, 1959;
Schwabe et al., 2012). In fact, only 20 – 60 % of the information given in oncological consultations is remembered by patients (Van Osch et al., 2014). To make well-informed decisions about treatment and to cope with their disease it is important patients receive and understand all information the oncologist is providing (Gabrijel et al., 2008; Sep et al., 2014). Therefore, it is important to reduce stress during oncological consultations to improve
patients’ recall.
Various studies have shown that oncologists’ communication style has an effect on patients’ stress, satisfaction and quality of life (Reblin et al., 2012; Sep et al., 2014; Vogel, Leonhart, & Helmes, 2009). A communication style that strengthens patients’ trust might be optimal for that purpose (De Haes & Bensing, 2009). Trust leads to less arousal and distress, improved information exchange and decision making, better adherence to medical advice, less second opinion seeking and higher satisfaction with the physician (Hillen et al., 2014; Sep et al., 2014; Van Osch et al., 2014). Hillen and colleagues (2014) showed that cancer patients’ trust is strengthened by a communication style that conveys competence (medical skills), honesty (telling the truth) and caring (involvement and sympathy). First, patients’ perception of oncologist’s competence can be expressed, for example, by emphasizing specialization in a particular operation (Hillen et al., 2014). Second, honesty of the oncologist is conveyed, for example, by indicating the wish to honestly inform the patient about the possible
complications (Hillen et al., 2014). Third, the oncologist can express caring by, for example, indicating availability to the patient in case of further questions (Hillen et al., 2014).
The present study was designed to examine if high trust-evoking communication leads to stress reduction and results in higher recall compared to a standard communication style. Results of a recent study by Sep and colleagues (2014) indicated that affective communication might reduce physiological arousal and increase information recall in bad news consultations. The present study will build on these results with some differences. First, as explained and
hypothesized by Sep and colleagues (2014), arousal and recall could be associated (the inverted U-shape function and the attention narrowing hypothesis). If the communication style of oncologists could lead to a decrease in arousal, this could influence information recall in a positive way. This association was not directly investigated by Sep and colleagues
(2014). Second, to increase generalizability, this study will focus on a gender neutral condition (colon cancer) and therefore includes both women and men as participants. In the study of Sep and colleagues (2014) only women participated. Third, in the current study psychophysiological measures will include a cardiovascular (heart rate) measure in addition to the measure of electrodermal activity that Sep and colleagues (2014) used, to create a more complete picture. Fourth, in addition to tonic psychophysiological measures (indicating arousal), we will also asses phasic electrodermal activity (i.e. Skin Conductance Responses, SCRs), indicating attention. Fifth, instead of manipulating oncologists’ communication style by adding empathic and reassuring remarks, the current study will specifically manipulate the level of trust that oncologists’ evoke, by adding three types of remarks exposed by Hillen and colleagues (2014): competence, honesty and caring. Sixth, the oncology consultation in the current project is not a bad news consultation, but a consultation about treatment. Previous studies (e.g. the study of Sep et al., 2014) contain bad news consultations, not treatment consultations. Treatment consultation also include important and stressful information (e.g. side effects) and occur frequently. We do not know yet how much arousal is provoked by treatment consultations which is why the current study includes these types of consultations.
In the present study, video vignettes of oncological treatment consultations will be used. These vignettes have been developed by Hillen and colleagues (2014) and contain a scripted oncological consultation between a patient and an oncologist. Scripted videos provide the opportunity to vary and study specific elements of communication (Van Vliet et al., 2012). Analogue patients (APs) watch the video vignettes while imagining that they are having the
consultation with the oncologist (Blanch-Hartigan, Hall, Krupat, & Irisch, 2013). APs are patients and/or healthy subjects who watch videotaped consultations while putting themselves in the shoes of the video-patient (Van Vliet et al., 2012). To examine the effects of
communication in an oncological setting, the use of video vignettes instead of real
consultations and APs instead of real patients has several advantages (Van Vliet et al., 2012). First, it enables standardization of certain elements, i.e. the oncologist, the visit characteristics and the communication (Van Vliet et al., 2012). By standardization, causal relationships can be examined in which effects cannot be attributed to these elements. Furthermore, it provides a higher chance of the required amount of participants instead of being dependent of patient flows (Van Vliet et al., 2012). Finally, it provides an alternative and ethical manner of
examining communication style during oncological consultations instead of manipulating the oncologist’s behavior which could possibly lead to suboptimal communication for patients (Blanch-Hartigan et al., 2013; Van Vliet et al., 2012).
Therefore, in this study we aim to test experimentally whether high trust-evoking
communication is associated with lower (analogue) patients’ psychophysiological arousal and improved recall of information provided in oncological consultations. Based on the literature, it is expected that a high trust-evoking communication style is negatively associated with extent of the psychophysiological response (Hypothesis 1) and positively associated with the recall of information (Hypothesis 2). Moreover, we aim to test if an association exists
between communication style, physiological response and recall by testing if the relationship between trust and recall is mediated by physiology (Hypothesis 3).
Method Design
This study used a randomized experimental design with two video-vignettes versions. The video vignettes were developed by Hillen and colleagues (2014). Eight different versions
of a video of an oncological consultation were created by Hillen and colleagues (2014), two of which were used in the present study. The two videos were identical in most respects: the oncologist in the video vignettes was a man, as a male oncologist still is the stereotype that patients hold of surgeons (Hillen, Van Vliet, De Haes, & Smets, 2013). Furthermore, a voice-over introduced the video. During the introduction the voice-voice-over explained the situation and gave instructions to the APs. Also, images of the setting were shown to make sure the APs got involved and got used to the setting. Finally, a patient perspective was used. This means that the camera was directed at the oncologist for most of the time and at the patient only for a short period of time. The type of cancer chosen for these video vignettes was colon cancer. This type of cancer was chosen because its incidence is comparable for men and women (Hillen et al., 2014). The video vignettes included the following elements: 1) summary of diagnosis and test results, 2) explanation what the surgery entails, 3) discussion of side effects and 4) time for questions of the patient. The duration of the standard video was 5 minutes and 29 seconds, and of the high-trust evoking video was 7 minutes and 28 seconds.
The contents of the two versions of video vignettes were identical because of the basic script that had been developed for the video vignettes (Hillen et al., 2013). The standard video vignette showed a doctor who informs the patient neutrally. In the high trust-evoking video vignette, the oncologist evoked trust by conveying competence, honesty and caring (Hillen et al., 2014). Demonstration of competence is being manipulated by the oncologist 1)
demonstrating to be up-to-date on the recent research literature, and 2) emphasizing his specialization in, and extensive experience with, the particular operation (Hillen et al., 2014). Honesty is manipulated by the oncologist 1) emphasizing his wish to exhaustively inform the patient, 2) indicating his wish to honestly inform the patient about the possible complications and 3) stressing his inability to rule out metastases with 100% certainty (Hillen et al., 2014). Caring behavior is demonstrated by the oncologist by 1) exploring a patient cue regarding
worry about possible metastases, 2) exploring the patients’ concern about a possible
colostomy, and 3) indicating availability to the patient in case of further questions (Hillen et al., 2014) (See Appendix A).
For each version of the video vignettes, two identical variations were used: one with a male patient and one with a female patient. These variations were made to match APs’ gender to optimize identification with the video patient.
Participants (APs)
The participants, i.e. APs, were Dutch students between 18 and 35 years old. A requisite to participate in this experiment was that the APs were healthy, not treated by a medical specialist and did not have any close relatives who are being treated for cancer. Before watching the video, the APs were asked to imagine that they were having the consultation with the oncologist (Blanch-Hartigan et al., 2013). Several studies demonstrated that APs can be included as proxies for real clinical patients to provide knowledge on physician-patient communication (Blanch-Hartigan et al., 2013; Van Vliet et al., 2012).
APs were randomly assigned to one of the conditions (standard communication style or high trust-evoking communication style). In order to allow the exclusion of unreliable data, the current study aimed to recruit at least 80 APs. Eventually, 97 APs were recruited to participate in the current study; the standard condition consisted of 48 APs and the high trust-evoking condition consisted of 49 APs. EDA data of three APs was excluded from analysis due to a not usable signal. Multiple reasons might explain this unusable signal, like cold fingers or disruption of the signal. For HR, data of thirteen APs was excluded due to a not usable signal and one since it was an extreme outlier. Including this outlier in the analysis resulted in an 1-point increase of standard deviation (SD) for the high trust-evoking condition, which led to the decision to exclude this outlier from analysis. Regarding recall, data of one AP was excluded because the recall questionnaire was not completed.
Measurements
Background characteristics. APs’ background characteristics were acquired by
completing a questionnaire before watching the video vignette (T0), assessing age, gender, educational level and medical knowledge (e.g. one item: “How much medical knowledge do you have?”; range: 1 = no knowledge at all and 7 = a lot of knowledge) (See Appendix B).
Engagement and credibility. Engagement of APs in the video was measured using the
Video Engagement Scale (VES). This questionnaire consisted of fifteen items (e.g. one item: “During viewing I was fully concentrated on the video”; range: 1 = fully disagree and 7 =
fully agree). In addition, three items to check credibility of the video vignettes were added to
this questionnaire (e.g. one item: “I thought the video was realistic”; range: 1 = fully disagree and 7 = fully agree). Engagement and credibility were checked using a questionnaire APs received after watching the video vignette (T1) (See Appendix C).
Manipulation check.To validate the effect of the manipulation of oncologist’s
communication style, the Trust in Oncologist Scale (TiOS), consisting of eighteen items, was used (e.g. one item: “This doctor is very careful and precise”; range: 1 = fully disagree and 5 = fully agree). In addition, three items were used to check the three components that evoke trust; competence, honesty and caring (e.g. one item: “This doctor is very honest”; range: 1 =
fully disagree and 5 = fully agree). The manipulation of trust was checked using the
questionnaire APs received after watching the video vignette (T1) (See Appendix C).
Electrodermal activity (EDA). EDA was measured by using a Biopac MP150 system
connected to a Windows 7 operating computer that was running the data acquisition program Acqknowledge 4.3 (Biopac). To measure EDA at 1000 Hz, the wireless BioNomadix EDA system (BN-PPGED) was used with two disposable electrodes, which were attached to the middle phalanx of the index and middle finger of the right hand with isotonic gel (type: EL507, Biopac). By letting the participants watch the video by themselves and giving them
the instruction to sit as still as possible, the occurrence of random responses due to e.g. movement or talking was minimized. The mean level of EDA was used as indicator of
psychophysiological arousal and is called Skin Conductance Level (SCL) in the analysis. The SCL was calculated for baseline and the video vignette. In addition to SCL, Skin Conductance Responses (SCRs) was measured as an indicator of phasic activity, thus attention. SCRs was calculated as the mean number of peaks per minute (threshold 0.05 micro Siemens). Like SCL, SCRs was calculated for baseline and the video vignette.
Heart rate (HR). HR was measured by using the Biopac MP150 system connected to a
Windows 7 operating computer that was running the data acquisition program Acqknowledge 4.3 (Biopac). To measure HR, the wireless BioNomadix ECG system (BN-RSPEC) was used with two disposable electrodes, which were attached to the chest with isotonic gel (type: EL503, Biopac). The average HR per minute was calculated from the filtered ECG signal and was calculated for baseline and the video vignette.
Recall. Recall was measured using both essay questions and multiple choice questions to
measure active recall and recognition (passive recall). We developed a 22-item questionnaire including 11 essay questions and (the same) 11 multiple choice questions with three answer options. These questions were developed after watching the video vignettes repeatedly and focusing on the most important information the oncologist provided. A pilot study was performed to test this recall questionnaire for ceiling effects and variety of questions. The maximum score of active recall in the pilot test was 22 and the maximum score of recognition was 11. Table 1 summarizes the descriptive statistics for the pilot test. Based on these results, we decided to adjust three questions because they were not understood properly or were too similar. Also, one question was added to ensure the questionnaire consisted of ten questions which facilitated scoring.
Table 1
Descriptive Statistics for the Pilot Study of the Recall Questionnaire
Variable Condition Min Max Mean SD
Active Recall Recognition 1 2 1 2 10.5 14.0 8.0 9.0 16.0 18.0 11.0 11.0 13.8 15.8 9.8 10.3 2.0 1.7 1.0 .8
Note. Condition 1 (n = 6) represents the condition in which the standard
communication style is used and condition 2 (n = 6) represents the condition in which the high trust-evoking communication style is used.
The final questionnaire used in the current study consisted of twenty items (See Appendix D). First, APs were asked to answer ten essay questions to assess active recall. Second, they were asked to answer the same ten questions but then they were provided three answer options to assess recognition. To score active recall, three categories were used; fully correct answer (2 points), partly correct answer (1 point) and incorrect answer (0 points). A total of 20 points could be obtained for active recall. To score recognition only two categories were used; correct answer option (1 point) and incorrect answer option (0 points). A total of 10 points could be obtained for recognition.
Procedure
APs signed up for the experiment via the website of the University of Amsterdam, after which they received an e-mail which included information about the experiment and in which their age and medical situation was requested. If they were eligible to participate, two
appointments were made; one for the experiment at the research location and one for the telephone call that occurred 24-28 hours after the experiment. To avoid that APs intentionally
memorized information from the video vignette, they were not told the telephone call
included a memory test. APs were told that this phone call was to evaluate the experiment and to make sure the AP would not be having emotional distress after this experiment.
During the experiment the APs first completed a baseline questionnaire (T0). After that, they were connected to the physiology equipment and watched a nature documentary to generate a baseline for the EDA and HR. Before watching, the APs were instructed to minimize movements, as this may affect psychophysiological measurement. Then, the APs watched one randomly selected version of the two video vignettes (the standard or the high trust-evoking). After watching the video, the APs completed the second questionnaire (T1).
24-28 Hours after the experiment the APs were telephoned to measure their information recall. Afterwards, a short debriefing took place to fully inform the APs about the study purposes, the APs were thanked for participation and they received a financial compensation or a study point.
Statistical analysis
All statistical analysis were performed using SPSS Statistics 21. At first, t-tests were used to: asses differences in background characteristics of APs; check engagement and credibility of the two video vignettes, and; perform a manipulation check. To analyse the individual physiological responses, we used three 2x2 repeated measures ANOVA’s. The condition (standard and high trust-evoking) was the between-factor and time (baseline and vignette) the within-factor. To test for differences in recall score between conditions, two t-tests were performed for active recall and recognition, respectively. Finally, a mediation analysis was performed to see if the relation between condition and recall is mediated by physiological response. A series of three regression analysis was performed to test whether this relation is significant.
Results Background characteristics
The two conditions, consisting of respectively 48 and 49 participants, are similar on background characteristics. Table 2 summarizes participants’ background characteristics.
Table 2
Background Characteristics of Participants Summarized per Condition
Variable Condition High trust-evoking (n = 49) Md Standard (n = 48) Md p Age 21 Frequency (%) 22 Frequency (%) .142 p Gender 1 = man 2 = woman Educational level 1 = Middle school
2 = Higher general and pre-university education 3 = Higher education 4 = University 5 = Other Medical Knowledge 1 = No knowledge at all 8 (16.3) 41 (83.7) 1 (2.0) 10 (20.4) 6 (12.2) 30 (61.2) 2 (4.1) 1 (2.0) 6 (12.5) 42 (87.5) - 12 (25) 3 (6.3) 33 (68.8) - 4 (8.3) .288 .751 .697
2 3 4 5 6 7 = A lot of knowledge 13 (26.5) 12 (24.5) 9 (18.4) 12 (24.5) 2 (4.1) - 8 (16.7) 15 (31.3) 5 (10.4) 15 (31.3) 1 (2.1) -
Engagement and credibility
An independent-samples t-test was conducted to compare APs’ engagement scores on the VES between the two video vignettes. Regarding the extent of APs engagement, there was no significant difference between the two video-vignettes versions, t = -.906, p = .367 (two-tailed). APs did not significantly feel more engaged in the video in which the high trust-evoking communication style is used (M = 71.2, SD = 13.2), compared to the video in which the standard communication style is used (M = 68.5, SD = 15.4). Also, there was no
significant difference in credibility between the two conditions, t = .197, p = .844 (two-tailed). Neither the video in which the high trust-evoking communication style is used (M = 14.9, SD = 3.4), nor the video in which the standard communication style is used (M = 15.1, SD = 3.8) was significantly perceived as a more credible oncological consultation.
Manipulation check
The trust-validation data was not normally distributed, which resulted in performing a Mann-Whitney U test instead of a t-test. Scores of the TiOS showed that APs in the trust-evoking condition reported more trust in the oncologist (M = 53.3, n = 49) compared to the standard condition (M = 44.6, n = 48), but this effect was not significant; U = 964, z = -1.531, p = .126, r = .155. In addition, APs in the high trust-evoking condition did not perceive the
oncologist as more competent (z = -.556, p = .578), honest (z = -.376, p = .707) or caring (z = -1.209, p = .227) than APs in the standard condition.
Electrodermal activity (EDA)
SCL. There was no interaction effect for SCL for the two conditions over time,
Wilks’ Lambda = 1.000, F(1,92) = .000, p = .989. This means that there was no difference between groups in their SCL-response to the video vignette. For time (baseline and video vignette), there was a significant effect, Wilks’ Lambda = .503, F(1,92) = 90.789, p = .000, indicating there was a significant change in SCL from the baseline to the video vignette. The main effect of condition was also significant, p = .026. As shown in Figure 1, APs in the high trust-evoking condition showed significantly higher SCL than APs in the standard condition during baseline as well as the video vignette.
Figure 1. Mean scores of SCL-response, measured in micro Siemens (µS), during the baseline
SCRs. There was no interaction effect for amount of SCRs, Wilks’ Lambda = .995, F(1,92) = .433, p = .512. This means that there was no difference between groups in their amount of SCRs to the video vignette. For time (baseline and video vignette), there was a significant main effect, Wilks’ Lambda = .653, F(1,92) = 48.893, p = .000, suggesting there was a significant change in SCRs from the baseline to the video vignette. The main effect of the two conditions was also significant, p = .016. As shown in Figure 2, this means that a significant difference between the standard condition and the high trust-evoking condition regarding the amount of SCRs both during baseline as well as during the video vignette.
Figure 2. Mean scores of SCRs , measured in micro Siemens (µS), during the baseline video
and the video vignette for the standard and the high trust-evoking condition.
Heart rate (HR)
There was an interaction effect for HR for the two conditions, Wilks’ Lambda = .942, F(1,81) = 4.500, p = .037. APs in the standard condition showed a larger increase of HR
between the baseline video and the video vignette than APs in the high trust-evoking
condition. Figure 3 shows the mean value of HR during baseline and during the video vignette for both conditions.
Because of the interaction effect, general interpretation of main effects was not appropriate, but results were reported. For time (baseline and video vignette), there was a significant effect, Wilks’ Lambda = .804, F(1,81) = 19.786, p = .000. The main effect of the two conditions was not significant, p = .769.
Figure 3. Mean scores of HR, measured in beats per minute (BPM), during the baseline video
and the video vignette for the standard and the high trust-evoking condition.
Recall
For both active recall and recognition scores, an independent-samples t-test was performed to assess whether recall differed significantly for the two conditions. The
difference in recognition scores was not significant (t = -.675, p = .502). APs in the high trust-evoking condition did not significantly show better passive recall, M = 8.9, than APs in the standard condition, M = 8.7 (88.5% correctly answered versus 87.3% in the standard
condition). However, the difference in scores of active recall was significant (t = -2.093, p = .039). APs in the high trust-evoking condition showed better active recall, M = 13.1, than APs in the standard condition, M = 11.9 (65.3% correctly answered versus 59.5% in the standard condition). There has been an increase of active recall of 5.9% in the high trust-evoking condition compared to the standard condition. Figure 4 shows the mean scores and the distributions of active recall scores for the two conditions.
Figure 4. Boxplot of active recall scores for the standard condition and the high trust-evoking
condition.
Mediation analysis
A mediation analysis was performed to test whether HR mediates the relationship between condition and active recall. To perform this mediation analysis, the difference score
of HR (formula: HR during video vignette – HR during baseline) instead of mean HR score is used. This way one value for HR is obtained to perform a mediation analysis using condition, difference score of HR and active recall score.
The mediation analysis was performed using the Baron and Kenny (1986) causal steps approach. The initial causal variable was condition, a categorical measure; the outcome variable was active recall, a continuous measure; and the proposed mediating variable was HR, also a continuous measure. Figure 5 shows this mediation model.
Figure 5. Mediation model.
Preliminary data screening suggested that there were no serious violations of the assumptions of normality. All coefficients reported here are unstandardized, unless otherwise noted; α = .05 two-tailed is the criterion for statistical significance.
The total effect of condition on active recall was significant, c = 1.167, t(94) = 2.093, p = .039; the high trust-evoking condition caused approximately a 1-point increase in active recall compared to the standard condition. Condition was significantly predictive of the hypothesized mediating variable, HR, a = -1.281, t(81) = -2.121, p = .037. When controlling for condition, HR was not significantly predictive of active recall, b = -.206, t(80) = -1.906, p = .060 (two-tailed). The estimated direct effect of condition on active recall, controlling for conflict, was c’ = .952, t(79) = 1.546, p = .126.
Active recall was not significantly predicted by condition and HR, with adjusted R² = .048 and F(2,79) = 3.044, p = .053.
The indirect effect, ab, was -.167. This was judged to be not statistically significant using the Sobel (1982) test, z = 1.418, p = .156. In this case, only the a and c coefficients were significant, the b coefficient and the ab product were not significant. By these criteria, the indirect effect of condition on active recall through HR was not statistically significant. Therefore, it can be concluded that the effects of condition on active recall were not mediated by HR. Figure 6 shows the path coefficients for this mediation analysis.
Figure 6. Mediation model with path coefficients.
Discussion
This study tested the effect of oncologists’ trust-evoking communication style on (analogue) patients’ psychophysiological response and their recall of information. As
expected, oncological consultations caused psychophysiological arousal in analogue patients (APs). The first hypothesis we predefined, implied that a high trust-evoking communication style is negatively associated with extent of the psychophysiological responses. This means a high trust-evoking communication style causes a smaller increase of APs’
psychophysiological activity compared to a standard communication style. The results of the current study confirmed this hypothesis for heart rate. The second hypothesis implied that a high trust-evoking communication style is positively associated with recall of information. Meaning that APs show higher recall after watching a high trust-evoking communication-style video vignette compared to a standard communication communication-style. The results of the current
study confirmed this hypothesis for active recall. The third hypothesis implied that
physiological response mediates the relationship between communication style and recall. However, the results of the current study did not confirm this hypothesis.
The results of the current study show an effect of oncologists’ communication style on heart rate (HR). APs showed a smaller increase in HR if the oncologist used a high trust-evoking communication style. This suggests that stress could be reduced by a communication style that evokes trust in patients. The effect on HR that has been found, brings forward new evidence regarding the stress responses that oncological consultations evoke. Previous studies already demonstrated an effect of communication on physiological response (Reblin et al., 2012; Sep et al., 2014), but the current study is the first to examine both HR and
electrodermal activity (EDA). As a result of the current study, not only EDA has now been demonstrated to be affected (Sep et al., 2014), also HR turns out to be affected by oncologists’ communication style. However, the current results do not show an effect of oncologists’ communication style on EDA, both Skin Conductance Level (SCL) and Skin Conductance Responses (SCRs). Contrary to what was expected, the high trust-evoking communication style evoked a similar increase in both SCL and SCRs as the standard communication style. Finding an effect for HR and not for EDA is remarkable because they are both under control of the autonomic nervous system which means they appear unconscious and are not
influenced by somebody’s will (Hasser & Moffitt, 2001). An increase of HR and EDA is caused by arousal or acute stress (Hasser & Moffitt, 2001). The effect of communication on HR and not on EDA does imply that they actually differ from each other. There are multiple possible explanations for the current findings. First, a higher sensitivity of HR on personal threats can explain why there has been found an effect for HR but not for EDA (Bradley, Codispoti, Cuthbert, & Lang, 2001). Second, EDA is influenced only by sympathetic activity whereas HR is influenced by both sympathetic and parasympathetic activity (Cacioppo,
Tassinary, & Berntson, 2007). Therefore, the current findings might be the result of the influence of the parasympathetic nervous system mostly, which led to an effect for HR and not for EDA. Third, the effect of condition that has been found for EDA might be another plausible explanation. Despite randomization of the participants, it has been found that the two conditions differed significantly during the baseline. The participants in the high trust-evoking condition showed a significant higher EDA than participants in the standard
condition. Therefore, the current study might not have been able to determine the effect of a high trust-evoking communication style on EDA. Fourth, the current study used treatment consultations instead of bad news consultations, that have been studied so far (Sep et al., 2014; Van Osch et al., 2014). The psychophysiological stress responses that occur during treatment consultations may differ from the psychophysiological stress responses that have been demonstrated so far.
Results of the current study not only show an effect of oncologists’ communication style on psychophysiological response, but also on cognitive performance. The use of a high trust-evoking communication style led to improved active recall of information. These findings reaffirm the effect of communication style on recall, which was also found in previous studies (Sep et al., 2014; Van Osch et al., 2014). Although the effect of
communication style on active recall is modest, it is remarkable that adding only a few trust-evoking elements in oncologists’ communication style results in an improvement of memory functioning, while the basic script of information did not differ between the two conditions. This means communication style has an important influence on cognitive performance and the recall of information from oncological consultations is improved when the oncologist uses a high trust-evoking communication style. The results of the current study only show an effect of communication style on active recall, not on recognition. This might be due to the fact that the group of participants was homogenous and consisted almost entirely of students who are
trained to perform on exams, including multiple choice questions. This may have influenced the obtained data. Thereby, the fact that we found an effect for active recall but not for recognition might be due to ceiling effects regarding the latter. This means the recall
questionnaire was not able to produce meaningful variability in scores of recognition, because the mean score was already close to the maximum score (Keeley et al., 2013; Stucki, Stucki, Brühlmann, & Michel, 1995; Ward, Guthrie, & Alba, 2014). When using a longer
questionnaire where the variability of score is more distributed, a significant effect for recognition may also be found when replicating the current study.
Although effects of oncologists’ communication style on both psychophysiological response (HR) and cognitive performance (active recall) have been found, no significant mediation effect of HR on the relationship between communication style and active recall was shown in the current study. The effect of HR on active recall was not significant, so the
intended relationship between arousal and cognitive performance, also called the inverted U-shape function of stress (Salehi et al., 2010), has not been confirmed. Nevertheless, a p-value of .060 may suggest some effect of HR on active recall. Less conservative tests (for example one-tailed) could have found a significant effect of HR on active recall. In addition, active recall is not statistically significantly predicted by a model which contains both condition and HR, but again a p-value of .053 suggests that this could be a reasonable prediction model. Moreover, adding HR as a predictor for active recall ensured that the previous significant effect of condition on active recall was no longer significant. This might be due to partial mediation. The relatively small effects of condition on active recall and on HR might be a possible explanation why the mediation analysis did not reach statistical significance. In sum, although no significant mediation effect of HR on the relationship between communication style and active recall was shown in the current study, the idea of (partial) mediation does not seem to be unrealistic based on these results.
Several strengths have contributed to the findings of the current study. The main strength is the strong design of the current study. The experimental design that was used, allowed us to: manipulate communication style of the oncologist; explore the specific effects of the different communication styles and; determine possible mediation. Another strength is the use of treatment consultations instead of bad news consultation as used in previous studies (Sep et al., 2014; Van Osch et al., 2014). The current study is the first to use these type of consultations and demonstrated that not only bad news consultations evoke stress responses and result in limited cognitive performance, but so do treatment consultations. Therefore, it seems not only important to improve communication in bad news consultations but also in treatment consultations. Also, taking both HR and EDA in account as a psychophysiological measures, as an indication of both arousal and attention, is a strength of the current study. This outlines a complete picture because the two unconscious stress responses are
supplementary. Further, performing a pilot-test for the recall questionnaire is a strength of the current study. This allowed us to check the items of the questionnaire and to adjust and add a few items to create a more valid recall questionnaire.
Besides strengths, the current study also has some limitations. First of all, instead of a sufficient amount, only a few men participated in the current study which makes it doubtful to say the conclusions are drawn of a mixed population. Second, in the current study we did not find a manipulation effect of trust between the two video vignettes, which might be caused by the use of students between 18 and 35 years old as APs. The students participated for either money or study points which may have influenced their motivation to complete the
questionnaires conscientiously. This might be a possible explanation for the deviating results compared to the previous study using these video vignettes (Hillen et al., 2014). Moreover, the students did not have any experience with oncological consultations. Participants in the study of Hillen and colleagues (2014) were cancer patients who might find it easier to
imagining themselves in the video vignette and to evaluate the vignette because they have got some experience with these consultations. Lastly, as a limitation, the active recall scores in the current study might be influenced a bit by one item, item five, on the recall questionnaire (See Appendix D). Prior to the experiment we compiled three answer options for the essay
questions (0, 1 or 2 points) to facilitate scoring during the telephone call. While scoring the actual answers for item five we experienced a variety in answers that could be scored for 1 point. We found that it was possible to gain 1 point by guessing. In this way, the guessed answers that included elements of the answer option we compiled, resulted in 1 point while other guesses did not lead to 1 point. Possibly, the use of a questionnaire without any chance of guessing rightly may have led to an even clearer effect of communication style on active recall.
From the results of the current study it can be concluded that using a high
trust-evoking communication style in oncological consultations reduces (analogue) patients’ stress responses and increases patients’ recall of information. Therefore, it might be useful to include high trust-evoking communication in training for medical students who will become future oncologists. However, to make plausible recommendations for these training sessions, more research is required. A recommendation for future research is to use a larger and more diverse population of APs, for example for age and gender. Not only students or cancer patients should be involved in further research (Hillen et al., 2014). This way it will be possible to draw conclusions for a mixed population. Also, it is recommended to conduct more research to investigate the relationship between psychophysiological response and recall and to specifically determine the relationship between communication style,
psychophysiological response and recall. Further, it is recommended to perform a study to check the manipulation of trust in the video vignettes that has been developed by Hillen and colleagues (2014) once again. The study must also include students (18-35 years old) to check
whether this group is actually able to detect the differences in communication style shown in the video vignettes. In this design, both versions of the video vignettes should be shown to the participants in which half of the group watch the trust-evoking video first, after which the standard video is shown, as the other half watch the two video vignettes in reversed order, to avoid order-effects. After each video vignette, APs should complete the manipulation check. Finally, it seems also important to use treatment consultation in future research. Previous studies (Sep et al., 2014; Van Osch et al., 2014) only used bad news consultations and concluding from the current study it might be relevant to clearly determine the effects of communication style on stress responses and cognitive performance in treatment consultations also. This might lead to an improvement of communication during this type of consultations and might ensure a reduced level of stress in cancer patients during out-patient consultations and a higher recall of information afterwards.
Over all, the current study is a strong addition in research on physician-patient communication. We confirmed the impact of oncological consultations on stress responses and recall of patients. Besides, we pointed out the positive effects of oncologists’ trust-evoking communication style. The results provide perspective for training development for medical students and for future research.
References
Blanch-Hartigan, D., Hall, J. A., Krupat, E., & Irisch, .T. (2013). Can naïve viewers put themselves in the patients’ shoes?: Reliability and validity of the analogue patient methodology. Medical Care, 51(3), 16-21.
Bradley, M. M., Codispoti, M., Cuthbert, B. N., & Lang, P. J. (2001). Emotion and motivation I: Defensive and appetitive reactions in picture processing. Emotion, 1(3), 276.
Cacioppo, J. T., Tassinary, L. G., & Berntson, G. (Eds.). (2007). Handbook of
psychophysiology. Cambridge University Press.
De Haes, H., & Bensing, J. (2009). Endpoints in medical communication research, proposing a framework of functions and outcomes. Patient Education and Counseling, 74(3), 287-294.
Easterbrook, J. A. (1959). The effect of emotion on cue utilization and the organization of behavior. Psychological review, 66(3), 183.
Gabrijel, S., Grize, L., Helfenstein, E., Brutsche, M., Grossman, P., Tamm. M., & Kiss, A. (2008). Receiving the diagnosis of lung cancer: Patient recall of information and satisfaction with physician communication. Journal of Clinical Oncology, 26(2) 297-302.
Guillet, R., & Arndt, J. (2009). Taboo words: The effect of emotion on memory for peripheral information. Memory & Cognition, 37(6), 866-879.
Hasser, E. M., & Moffitt, J. A. (2001). Regulation of sympathetic nervous system function after cardiovascular deconditioning. Annals of the New York Academy of Sciences, 940(1), 454-468.
Hillen, M. A., De Haes, H. C. J. M., Stalpers, L. J. A., Klinkerbijl, J. H. G., Eddes, E. H., Butow, P. N.,…Smets, E. M. A. (2014). How can communication by oncologists enhance patients' trust? An experimental study. Annals of Oncology, 25(4), 896-901.
Hillen, M. A., Van Vliet, L., De Haes, H. C. J. M., & Smets, E. M. A. (2013). Developing and administering scripted video vignettes for experimental research of patient–provider communication. Patient Education and Counseling, 91(3), 295-309.
Jansen, J., Van Weert, J. C. M., De Groot, J., Van Dulmen, S., Heeren, T. J., & Besing, J. M. (2010). Emotional and informational patient cues: the impact of nurses’ responses on recall. Patient Education and Counseling, 79(2), 218-224.
Keeley, J. W., English, T., Irons, J., & Henslee, A. M. (2013). Investigating halo and ceiling effects in student evaluations of instruction. Educational and Psychological
Measurement,73(3), 440–457.
Kreibig, S. D. (2010). Autonomic nervous system activity in emotion: A review. Biological
psychology, 84(3), 394-421.
Reblin, M., Ellington, L., Latimer, S., Uchino, B. N., Roter, D., & Maxwell, A. (2012). Communication Style Affects Physiological Response in Simulated Cancer Risk Reduction Interactions. Journal of Applied Biobehavioral Research, 17(2), 129-156. Salehi, B., Cordero, M. I., & Sandi, C. (2010). Learning under stress: The inverted-U-shape
function revisited. Learning & Memory, 17(10), 522-530.
Schwabe, L., Joëls, M., Roozendaal, D., Wolf, O. T., & Oitzl, M. S. (2012). Stress effects on memory: An update and integration. Neuroscience & Biobehavioral Reviews, 36(7), 1740-1749.
Sep, M. S. C., Van Osch, M., Van Vliet, L. M., Smets, E. M. A., & Bensing, J. M. (2014). The power of clinicians’ affective communication: How reassurance about non-abandonment can reduce patients’ physiological arousal and increase information recall in bad news consultations. An experimental study using analogue patients.
Stucki, G., Stucki, S., Brühlmann, P., & Michel, B. A. (1995). Ceiling effects of the Health Assessment Questionnaire and its modified version in some ambulatory rheumatoid arthritis patients. Annals of the Rheumatic Diseases, 54(6), 461-465.
Van Osch, M., Sep, M. S. C., Van Vliet, L. M., Van Dulmen, S., & Bensing, J. M. (2014). Reducing patients’ anxiety and uncertainty, and improving recall in bad news consultations. Health Psychology, 33(11), 1382.
Van Vliet, L. M., Van der Wall, E., Albada, A., Spreeuwenberg, P. M. M., Verheul, W., & Bensing, J. M. (2012). The validity of using analogue patients in practitioner–patient communication research: Systematic review and meta-analysis. Journal of General
Internal Medicine, 27(11), 1528-1543.
Vogel, B. A., Leonhart, R., & Helmes, A. W. (2009). Communication matters: The impact of communication and participation in decision making on breast cancer patients’ depression and quality of life. Patient Education and Counseling, 77(3), 391-397. Ward, M. M., Guthrie, L. C., & Alba, M. (2014). Dependence of the minimal clinically
important improvement on the baseline value is a consequence of floor and ceiling effects and not different expectations by patients. Journal of Clinical Epidemiology, 67(6), 689-696.
Appendix A
Overview of script additions for the manipulation of trust Competence
1. The scientific literature clearly demonstrates that [this operation] offers the best chance at survival. I recently attended a scientific conference, where it was confirmed once more that this operation is the best possible treatment.
2. We perform [this surgery] very often. In our center we all have our specialization, so that we perform certain procedures much more often to become even more skilled in them. So I perform this operation a lot.
Honesty
1. [The risk of a colostomy] is something I really want to discuss with you, even though the chances of it are slight. Because I believe that you as a patient have the right to be informed about such small risks, to avoid that you will be taken unaware by this afterwards.
2. All these [complications] will not necessarily occur, but I do want to discuss them with you. Because, even though we do our very best, it remains hard to predict: we can never rule out that you will experience complications. This way at least you are well informed.
3. I should add that we can never be 100% sure [that the cancer will be completely removed]. So I can never give you an absolute guarantee. There is always a slight chance that we run into an abnormality during surgery. But for now all signs suggest that you will be cancer-free after the operation.
Caring
1. Oncologist: Were you very worried about it [the result of the scan]?
Patient: Yes, in a way…After all, it is your biggest fear in such a moment. But it seems to have turned out well, hasn’t it?
Oncologist: Yes, it seems so. So that is good, and at least one thing we don’t have to worry about anymore.
2. Oncologist: Do you particularly dread the thought of a colostomy?
Patient: Yes, it seems horrible, this bag attached to your stomach which may open at any moment, with all that filth…And you can’t show yourself anywhere anymore…
Oncologist: Yes, I understand what you mean.
3. But I will talk to you before the operation in any case. And it is important to know that I am always available for you in case of questions. This is our phone number. I may not always answer the phone myself, but if necessary I will return your call when I do have the time.
Appendix B
Overview of T0 questionnaire
Background characteristics
What is your gender? What is your age in years?
What is your highest level of education? How much medical knowledge do you have?
Appendix C
Overview of T1 questionnaire
VES
1. During viewing I was fully concentrated on the video.
2. During viewing it was as if I was present at the events depicted in the video. 3. When I was viewing the video, my thoughts were only with the video.
4. After the video was finished, I had the feeling I came back into the 'real' world. 5. When I had been viewing for a while, it seemed as if I had become the patient in my thoughts.
6. I empathized with the patient. 7. The video affected me.
8. When I was viewing the video, I was in the world of the video in my thoughts. 9. During viewing, I felt insecure when the patient felt insecure.
10. I found the video moving. 11. I felt for the patient.
12. During viewing, I was hardly aware of the space around me. 13. I had the feeling I went through what the patient went through. 14. In my imagination, it was as if I were the patient.
15. Because of the video, feelings arose in me.
Three items to check credibility
1. I thought the video was realistic. 2. I thought the video was credible.
TiOS
1. This doctor is very careful and precise.
2. This doctor is totally honest in telling you about the different treatment options available for your condition.
3. This doctor always gives you honest information about your prospects. 4. This doctor strongly cares about your health.
5. This doctor always tells you everything you want to know about your illness. 6. You think this doctor can handle any medical situation, even a very serious one. 7. This doctor always takes time with you.
8. This doctor explains everything so that you can consent to medical decisions. 9. Sometimes you worry that the doctor’s medical decisions are wrong.
10. This doctor only thinks about what is best for you.
11. Sometimes this doctor does not pay full attention to what you are trying to tell him/her. 12. The physician would always tell you the truth about your health, even if it was bad news. 13. You have doubts whether this doctor really cares about you as a person.
14. This doctor listens with care and concerns to all the problems you have. 15. This doctor will do whatever it takes to get you all the care you need. 16. This doctor is available for you whenever you need him/her.
17. You have no worries about putting your life in this doctor’s hands. 18. All in all, you have complete trust in this doctor.
Three items to check competence, honesty and caring
1. This doctor is an expert to me. 2. This doctor is very honest. 3. This doctor is very caring.
Appendix D
Overview of items of the recall questionnaire for both essay and multiple choice questions and the given multiple choice answer options
Essay and multiple choice questions Multiple choice answer options
1. There is found a malignant tumor. Can you describe the exact location?
2. What is the result of the CT scan?
3. What is for this patient the most obvious option as a treatment?
4. After how many weeks the temporary stoma can be removed with a small operation?
5. The physician indicated that it is necessary for the patient to receive information from the stoma care nurse. What kind of
information will the patient receive from her?
A. In the upper part of the esophagus B. In the lower part of the colon (correct) C. In the middle of the stomach
A. A metastasis to the liver B. No metastases (correct)
C. Multiple metastases throughout the body A. An operation in which the part of the colon where the tumor is located is removed (correct)
B. Chemotherapy
C. Irradiation to reduce the tumor A. After 4 to 10 weeks
B. After 6 to 12 weeks (correct) C. After 10 to 16 weeks
A. What it is like to live with a temporary stoma (correct)
B. How long the stoma must remain C. How other people experience life with a stoma
6. The doctor showed and told where the temporary stoma will be. Can you tell me the location?
7. The doctor mentioned four possible complications of the surgery. Can you name them?
8. In addition to the four possible
complications, the doctor emphasized the two standard complications of surgery. Can you name them?
9. How can a potential abscess be remedied?
10. With whom will the patient be mainly in contact for the coming period of time?
A. (Around) your belly button (correct) B. On your hip
C. On your back
A. An abscess in the abdomen, pain, nausea and a bleeding
B. Pain, a wound infection, an abscess in the abdomen and scar tissue
C. Risk of leaks, an abscess in the abdomen, a bleeding and a wound infection (correct) A. Pain and bleeding
B. Infection and bleeding (correct) C. Infection and scarring
A. By cutting it out B. With medication C. With a drain (correct) A. With the doctor B. With the intern
