The Association between Shared
Decision Making and Patients’
Trust in their Oncologists
An Observational Study
Masterthese
Irini Sorial
6072968
Programmagroep Klinische Psychologie
Faculteit der Maatschappij- en Gedragswetenschappen
Supervisoren:
dr. E. G. Engelhardt (LUMC/VUmc)
dr. M. A. Hillen (AMC)
The Association between Shared Decision Making and Patients’ Trust in their
Oncologists
ABSTRACT
BACKGROUND: In the medical oncological setting, breast cancer patients are often eligible for
adjuvant systemic treatment, i.e., chemo- and/or hormonal therapy, after having their tumor surgically removed. In many cases, there is no clear medically better option between these adjuvant treatments, so it is important that the oncologist involves the patient in choosing the kind of treatment that is most suitable, making use of shared decision making (SDM). SDM is a process between physician and patient to make health-related decisions together, based on the best available evidence and the patient’s goals and values. Implementing SDM could influence the level of trust patients have in their physician.
AIM: The aim of this study was to investigate whether there is a relationship between 1) observed SDM, 2) experienced SDM and 3) preferred SDM and the level of trust breast cancer patients have in their
oncologist. It was hypothesized that a greater level of SDM in consultations involving decisions about adjuvant systemic treatment, strengthens patients’ trust in their oncologist. By way of exploration, it was examined which patient and physician characteristics can be associated with increased trust.
METHODS: Audio recorded consultations of 101 breast cancer patients with their oncologist were
analysed. All consultations involved a decision about whether to undergo adjuvant systemic therapy and which. SDM was objectively scored by the researchers using the validated Observing Patient
Involvement In Decision-making rating scale (OPTION-12), and trust was reported by patients using the Trust in Oncologist Scale (TiOS). An adequate level of SDM was defined as score ≥ 50).
RESULTS: Patients’ trust in their oncologist was very strong, with an average score of 4.05 (SD = 0.56)
out of 5. Observed SDM scores ranged between 2.08 and 56.25 on a scale from 0 – 100 with a mean score of 15.5 (SD = 11.6). In 2 of the101 consultations an adequate level of SDM was reached.
Observed SDM was not associated with patients’ trust in their oncologists (Pearson’s R = .021; p = .837) and neither was experienced SDM in choosing hormonal therapy (F(1) = 0.257; p = .613) or
chemotherapy (F(1) = 0.079; p = .780). A one-way ANOVA showed no relation between TiOS scores and the level of preferred involvement in SDM (F(2) = .544; p = .582). The explorative multiple
regression analysis showed that a bigger size of the tumor (p = .035; β = .210) and oncologists’ use of the prediction model Adjuvant! during the consultation (p = .031; β = .215) were significantly associated with an increased level of patients’ trust in their oncologist.
CONCLUSION: SDM was not adequately implemented in the vast majority of consultations.
Nevertheless, patients’ trust in their oncologist was high. This finding suggests that application of SDM during the consultation is not associated with the level of trust. Future research needs to assess what higher levels of SDM during the consultation would mean for trust levels.
INTRODUCTION
As a famous English proverb goes, it is an equal failing to trust everybody, and to trust nobody. Trust is vital for human social interaction, since no trust would mean not being able to have faith in anyone. Hence, trust is needed in all different aspects of a society so it can function, for instance when going to the doctor and getting advice, but not being able to believe them since there is no trust. Therefore, trust is especially essential from a patient’s point of view, in a patient-physician relationship, since the
relationship between a patient and a physician is characterized by a knowledge and power imbalance, whereby the patient is dependent on the physician. Patients need to have trust in their physician as he or she has the ability to take care of their health, an ability patients can lack (Mechanic & Meyer, 2000). On the other hand, a physician needs the trust of a patient to convey treatment guidelines and augment compliance (Skirbekk, 2009). Thus, trust is of significant relevance for a therapeutic alliance between a patient and their physician and success of treatment (Ridd, Shaw, Lewis & Salisbury, 2009; Fuertes, Mislowack & Bennett, 2007).
Trust in a physician-patient relationship has received considerable research attention in the primary care setting. However, research devoted to trust in the clinical setting, and especially in the oncology setting, a large field in medicine, is more scarce. Cancer is a frequently occurring disease with 5% of the Dutch population getting it at one point in their life (http://www.cijfersoverkanker.nl/) and the importance of mutual trust during oncology clinical encounters of consultations is considered to be very high as there is a lot at stake (Hillen, De Haes & Smets, 2011). Cancer patients receive complex and life-threatening information, face difficult treatment decisions, and have no guarantees that the available treatment options will yield them any benefit. This makes them a vulnerable population that needs a significant amount of trust in their oncologist. The fact that in qualitative studies, ‘trust’ was often spontaneously put forward by patients, supports the idea that it is of importance to them (Hillen, Onderwater, Van Zwieten, De Haes & Smets, 2012). Moreover, trusting cancer patients worry less about treatment and are more likely to adhere to medical treatment and advice (Hillen et al., 2011). All in all, trust seems to be of vital importance in oncology.
The recent increase of patient autonomy, combined with global internet access and the change of health-care organization concerning insurance policies and less clarity on the costs of treatment, have led to the fear that patients’ unconditional trust in their oncologist might be eroding (Mechanic, 1996; McKinstry, Ashcroft, Car, Freeman, & Sheikh, 2006). A review by Hillen et al. (2011) investigating cancer patients’ trust in their physician has shown that patients, despite their increasing autonomy, still have substantial trust in their oncologists. The authors also concluded that patients trust physicians who they perceive as technically competent and honest. Patients also trust physicians who display facilitative behaviors and ways of communication, and physicians with whom a continuous relationship exists. Such trust is associated with a facilitation of the communication and medical decision-making process.
By addressing patients’ increasing need for autonomy, physicians may be able to prevent a potential erosion of trust, which still could be expected (McKinstry et al., 2006). One of the most discussed methods of promoting patient’s autonomy is shared decision making (SDM). SDM is the process by which the patient and the physician make health-related decisions together based on the best available evidence and the patient’s goals and values (Couët et al., 2013). The focus of SDM lies in the process of decision making rather than the decision itself, with consideration of patient’s goals and values as opposed to looking only at the facts and statistics. SDM has also been applied in oncology settings
(Couët et al., 2013). Stiggelbout, Pieterse and De Haes (2015) described four key characteristics of SDM. Firstly, patients are made aware that a decision needs to be made. Secondly the physician provides the patient with evidence-based medical information about the benefits and harms associated with all relevant treatment options. Thirdly, the patient’s values and preferences are explored, and lastly, an agreement is reached on the treatment to implement. SDM has been recommended as a key feature of good clinical practice by the World Health Organization (WHO, 1994). It permits patients to weigh the benefits and risks of a certain therapy with their physician and make reasoned informed choices
(Stiggelbout et al., 2012). On the other hand, SDM possibly enables patients to be more conscious about the uncertainty surrounding the information provided, prohibiting them to follow their doctor’s
decisions blindly. Subsequently, this might downgrade trust towards physicians. Nonetheless, recent qualitative studies have shown that a sense of SDM was particularly associated with the expression of higher trust in doctors among patients at general practice settings (Croker et al., 2013; Butterworth & Campbell, 2014). A recent review by Shay and Lafata in 39 studies found that SDM could lead to better affective-cognitive outcomes, defined as understanding, satisfaction and trust. Yet, the presence of SDM was measured by patient reports, being a subjective measure and not tallying with what happens during a clinical consultation (Shay & Lafata, 2015). The authors also concluded that the number of studies included that examined affective-cognitive outcomes was small, thus more research is needed in terms of patient affective outcomes, such as trust and satisfaction, and their relationship to SDM. Furthermore, the review did not investigate the effects of SDM on trust in an oncology setting and hence no solid evidence is available in cancer care.
More comprehensive research is necessary to increase the understanding of how physicians could contribute to trust, in light of the fear that this trust is fading. The common belief nowadays is that physicians who increase their patients’ autonomy successfully gain their trust. A possible and likely mechanism of physicians contributing to a patient’s autonomy and thus trust, is SDM. As mentioned above, no clear evidence exists on how and whether SDM increases cancer patient’s trust in their
oncologists. Such knowledge is essential, as a more trusting relationship between patients and physicians could eventually improve health care quality and outcomes.
Consequently, in this study, the association between the application of SDM by the oncologist and patients’ trust in them was investigated. This was done by examining consultations between oncologist and patient, which focused on the decision about potential further treatment, i.e., adjuvant systemic treatment. Adjuvant treatment is additional cancer treatment given after the primary treatment, i.e., surgery, to reduce the risk of the cancer returning. Systemic entails that the treatment affects the whole body. In this study, the options for adjuvant systemic treatment were: both chemotherapy and hormonal therapy, only chemotherapy or only hormonal therapy. In the consultations, there was also an option to decide on no treatment at all.
Three different forms of SDM were distinguished in this study, namely observed SDM, experienced SDM and preferred SDM. Observed SDM involves the formal rating of aspects of SDM by objective coders, and focused on the oncologist’s efforts to stimulate patient participation during the decision making consultation. Experienced SDM was reported by the patients and was based on how they perceived the interaction with their physician. Preferred SDM is the patient’s preferred level of
involvement in the decision making process. The necessity of making a decision during the consultation and the lack of a clear medically ‘best’ treatment option (i.e., the decision was preference-sensitive) made these consultations especially suitable to assess whether and how SDM was applied.
The consultations and data for this study were collected through the IBIS-study (Informeren en Beslissen In de Spreekkamer), a large multicenter observational study conducted with the aim to examine the role of a risk prediction model in conveying prognostic probabilities to breast cancer patients and assessing patients’ understanding of these probabilities (Engelhardt et al., submitted).
In short, the aim of this study was to investigate the relationship between SDM and trust between oncologist and patient. To achieve this, several research questions were formulated and tested:
1. Does the observed level of SDM predict patients’ level of trust in their oncologists? 2. Does the level of experienced SDM by the patient predict their level of trust in their
oncologist?
3. Does conformity between observed and preferred involvement of patients in medical decisions predict trust of patients in their oncologist?
4. Which patient and physician characteristics predict patients’ trust in their oncologists? 5. Which patient and physician characteristics predict patients’ level of involvement in
treatment decision making?
Based on available literature, it was hypothesized that the:
1. observed level of SDM predicts trust of patients in their oncologists.
2. level of experienced SDM by the patient predicts trust of patients in their oncologists. 3. conformity between observed and preferred involvement of patients in medical decisions
predicts trust of patients in their oncologists.
By way of exploration, it was examined which patient and physician characteristics: - predicted patients’ trust in their oncologists;
METHODS Participants
The participants - all female - were recruited for the IBIS-study. All patients in the IBIS-study had stage I-III breast cancer and were medically eligible to receive adjuvant systemic treatment with a curative intent (i.e., chemotherapy, hormonal therapy or both), after having their breast tumor surgically removed. Patients were enrolled prior to their consultation with their medical oncologist to discuss whether or not they would receive adjuvant systemic treatment.
A randomly selected sample of patients from the IBIS-study was included in this study. A power analysis using G*Power version 3.1.9.2 (Faul, Erdfelder, Buchner & Lang, 2009) was conducted to determine how many patients needed to be included in this study. As there were 12 variables (Table 3, see
appendix) in the prediction model in this study, the minimum sample size for a linear multiple regression to detect relevant differences with a power of .95, using an alpha of .05 and Cohen’s F2 = .15, was 89
subjects. The random sample size of 101 patients was chosen due to feasibility arguments, since scoring consultations is very labor intensive.
Procedure
In the IBIS-study, the consultations between patient and physician during which a decision had to be made regarding adjuvant systemic treatment, were audiotaped, transcribed, coded and analyzed. Additionally, a telephone interview was conducted with all participating patients and they received a questionnaire after the telephone interview.
Patients were recruited at eight university and general teaching hospitals in the Randstad region in the Netherlands. They were informed about the study before the consultation with their oncologist. The institutional review board of the participating hospitals approved the study protocol. After consent of a patient to participate in the study, the consultation between the oncologist and the patient was recorded. The audiotapes were transcribed verbatim by research assistants (with a background in medicine or health sciences), and the transcripts were reviewed for the current study by trained researchers (with a background in medicine, health sciences and psychology). Consultations were double coded by two researchers until adequate inter-rater reliability was achieved (Cohen’s Kappa ≥ 0.7).
Materials
The level of observed SDM during the consultation was assessed using the OPTION-12 rating scale, which stands for Observing Patient Involvement In Decision-making (Elwyn et al., 2005). The OPTION-12 is a discrete measure of SDM and is designed to measure the extent to which clinicians (medical, nursing or other relevant professional) involve patients in decisions within consultations. The items reflect communication behaviors needed to involve patients in the decision-making process, and the score reflects the clinicians’ competence. It has twelve items (see Table 1, appendix) to be scored on a 5-point Likert scale, ranging from 0 (the behavior is not observed) to 4 (the behavior is observed and executed to a high standard), which makes the minimum score 0 and the maximum sum score 48. The sum score was converted to a 0 - 100 scale in this study. Validity has been tested in multiple studies and was found to be sufficient (Elwyn et al., 2013). In the present study, inter-rater reliability was sufficient for all items, with Cohen’s Kappa ≥ 0.7. Internal consistency was also sufficient (Cronbach’s Alpha = .794). OPTION-12 was coded by the researchers in the IBIS study using the transcripts of the
audiotaped consultations. Couët et al. (2013) adopted a minimum score for adequate SDM of 50 (on the 0 - 100 scale), which was also maintained in this study. A mean score lower than 25 indicates that even a
‘perfunctory or unclear attempt to perform the behavior is not observed consistently, while a score higher than 25 but lower than 50 indicates that an attempt (perfunctory or unclear) is observed, but that the behavior itself could not be performed to a ‘baseline skill level’ (Couët et al., 2013).
Patients’ experienced SDM and their preferred level of involvement in treatment decision making were measured with a formally adjusted format of the Control Preferences Scale (CPS) (Degner & Sloan, 1992). The adjusted format that was used consisted of one open-ended question, namely: “Who made the final decision to start treatment X in your opinion?” (Sutherland, Llewellyn-Thomas, Lockwood, Tritchler & Till, 1989). This ordinal scale, which was extensively validated in the Western population, originally had five possible response-categories, namely: 1) patient-driven, 2) patient-driven with oncologist’s advice taken into account, 3) shared, 4) oncologist-driven and 5) oncologist-driven with patient’s wishes taken into account. In this study, patients were interviewed by telephone within seven days after the consultation to assess their experienced involvement in decision making. They were asked whom they felt had made the final treatment decision with aforementioned open-ended question. Patients’ answers were classified in three response categories instead of the original five, indicating a preference for either 1) patient-driven, 2) shared or 3) oncologist-driven decision-making. The answers were independently coded and categorized by two researches, with a sufficient inter-rater reliability: Cohen’s Kappa ≥ 0.7.
To obtain patients’ preferred level of involvement in treatment decision making, a pen and paper survey was used, after the consultation. Patients were asked to choose one of the three CPS categories that best reflected their preference.
Trust was investigated using the well-validated Trust in Oncologist Scale (TiOS), a self-report scale that measures trust by assessing four different dimensions, i.e., Competence, Fidelity, Honesty and Caring (Hillen et al., 2012). The TiOS consists of 18 items (Table 2, see appendix), scored on a 5-point Likert scale ranging from 1 (completely disagree) through 5 (completely agree), which makes its score a discrete measure. It was conducted after the telephone interview. The three reversely phrased items on the scale were first recoded. Then the scores on the answers to all 18 items were summed and divided by 18 to obtain an average trust score, with higher scores indicating higher trust. A high internal consistency was found (Cronbach’s Alpha = .929).
During some of the consultations physicians used the online software program Adjuvant! Online. This is a digital prediction model used as a decision support tool since it provides estimates of disease-free probabilities and specific breast-cancer survival rates with or without adjuvant systemic treatment within 10 years (Ravdin et al., 2001).
Statistical Analyses
Firstly, descriptive analyses on the variables were performed and continuous variables were presented as means with standard deviations (±SD). Dichotomous variables were presented as absolute numbers with percentages.
Research question (RQ) 1 was tested using Pearson’s correlation, to see if the observed level of SDM was associated to patients’ level of trust in their oncologists. It was hypothesized that the greater the observed level of SDM, the greater patients’ level of trust in their oncologists would be. Initially this would have been tested with a linear regression analysis, but due to the found descriptives which indicated very low OPTION-12 scores which meant that there was an insufficient variation to make a meaningful comparison with the TiOS, a correlation analysis was chosen.
RQ2 was tested using descriptive analyses and a one-way ANOVA, investigating the association between the level of experienced SDM as measured with the adjusted CPS scale and patient’s trust. Also for this
RQ a linear regression analysis was planned, but the ANOVA appeared more fit to test the hypothesis that the greater the level of experienced SDM, the greater the level of trust patients have in their oncologist, since the amount of experienced SDM seemed not sufficient to conduct a linear regression analysis.
To test RQ 3, namely whether congruence between patients’ observed and preferred involvement in medical decisions predict patients’ trust in their oncologist, the OPTION-12 scores (discrete variable, score range minimal = 0 and maximum = 100) were compared to the preferred level of involvement in decision-making as measured with the CPS scores (categorical variable, i.e., patient-driven, shared and oncologist-driven) using the Kruskal-Wallis test or a one-way ANOVA as appropriate. It was
hypothesized that conformity between observed and preferred involvement of patients in medical decisions yields higher scores of level of trust in the oncologist.
Explorative analyses were performed to answer RQ 4 and RQ 5. To answer RQ 4, a prediction model was built using linear regression analysis with backward selection. It was evaluated which baseline characteristics of oncology patients, their oncologists and SDM scale scores predicted patients’ trust in their oncologist as measured by the TiOS (dependent variable). The variables included in the model are listed in Table 3 (see appendix).
RQ 5 was not pursued in the end, considering the very low observed SDM scores and the lack of variability of the OPTION-12 scores, but a linear regression analysis with backward selection was planned to evaluate which baseline characteristics of oncology patients, their oncologists and SDM scale scores predicted patients’ level of involvement in treatment decision making.
Conducting analyses on complete cases only is known to be prone to biased effect estimates. Patients with more than two missing items on the TiOS were excluded. To correct for four patients who had one missing score on the TiOS, median imputation was used to attempt reducing the risk of bias. This was the case for TiOS-items 2, 5, 8 and 13.
A two-sided p-value < .05 was considered significant. Statistical analyses were performed using SPSS version 21.0 (Chicago, IL).
RESULTS
A total of 101 female patients with breast cancer were included in this study. All women had undergone an operation to remove their tumor prior to the consultation with their oncologist. Of the consultations, 94 (93%) were conducted by a specialist and 7 (7%) by a resident. Conducted by a male physician were 57 (56%) of the consultations. Other descriptive variables are displayed in Table 4 in the appendix. Results are presented per research question (RQ).
RQ 1. Does the observed level of SDM predict patients’ level of trust in their oncologists?
Hypothesis 1: The greater the observed level of SDM, the greater patients’ level of trust in their oncologists.
The observed level of SDM was not significantly correlated with the level of trust of the patient in their oncologist (Pearson’s R = .021; p = .837). Linear regression analyses could not be done as OPTION-12 scores were non-normally distributed, with skewness of 1.37 (SE = 0.24) and kurtosis of 1.74 (SE = 0.48). On a scale ranging from 0 - 100, OPTION-12 scores ranged between 2.08 and 56.25, with a mean of 15.5 (SD = 11.6), creating insufficient variation to allow for a meaningful comparison with the TiOS, since in only 2 of the 101 consultations the minimum adequate level of SDM (i.e., score ≥ 50) was reached.
RQ 2. Does the level of experienced SDM by the patient predict their level of trust in their oncologist?
Hypothesis 2: The greater the experienced level of SDM, the greater patients’ level of trust in their oncologists. Patients’ experience of SDM was measured using the adjusted format of the CPS, which was an open-ended question, asking who the patient thought had made the final decision concerning their treatment. The answers were placed into three categories (see Figure 1, appendix).
Of the patients that were eligible for chemotherapy, 66% felt like they had made the final decision. Of the patients that were eligible for hormonal therapy, 46% felt like they had made the final decision. Respectively 8% and 9% of the patients explicitly reported that a shared decision had been made. ANOVAS with trust as the dependent variable and the experienced level of SDM as the independent variable (for hormonal therapy and chemotherapy separately) showed no significant relations between TiOS scores and the three experienced SDM categories, namely whether the decision was made 1) by the oncologist, 2) by the patient or 3) together. Hormonal therapy: F(1) = 0.257; p = .613 and
chemotherapy: F(1) = 0.079; p = .780.
RQ 3. Does congruence between patients’ observed and preferred involvement in medical decisions predict patients’ trust in their oncologist?
There were only two consultations with an OPTION-12 score ≥ 50, which meant that meaningful statistical tests could not be done with this low-variation variable. Therefore, patients’ preferred level of SDM as measured with the adjusted CPS scale to test associations between preferred SDM and trust was used. The original hypothesis was adjusted and tested using a One-way ANOVA analysis.
Hypothesis 3 (adjusted): If patients’ preferred SDM was either patient-driven or shared, they will have lower TiOS scores than the patients who preferred an oncologist-driven decision, since practically all consultations were oncologist-driven, based on the very low amount of observed SDM.
A one-way ANOVA with trust as the dependent variable and the preferred level of SDM as the independent variable showed no significant difference in the average TiOS score between the three preferred SDM level categories, namely whether decision-making was 1) patient-driven, 2) shared, or 3) oncologist-driven (F(2) =0.544; p = .582).
Explorative analyses
RQ 4. Which characteristics predict patients’ trust in their oncologists?
To assess which variables influenced trust, we built a model using linear regression with backward stepwise selection. The full model consisted of the variables described in Table 2, i.e., age, education, having children, marital status, preferred SDM, hospital affiliation, gender oncologist, use of Adjuvant!, duration of the consultation, TNM stage of disease, size of tumor, and tumor-positive lymph nodes. Of these variables only tumor size (p = .035) and the use of Adjuvant! (p = .031) were associated with trust levels. Larger tumor size was associated with higher levels of trust, and if the oncologist’s explanation included probabilistic information from Adjuvant!, trust levels were higher (see Table 5 in appendix).
RQ 5. Which characteristics predict patients’ level of involvement in treatment decision making?
After evaluating the SDM scores in the study population, it was decided not to pursue this last research question. Considering the very low observed SDM scores (Table 4, see appendix) and lack of variability in OPTION-12 scores, this patient population is not suitable to answer this research question, as the SDM level was inadequate in nearly all consultations. There is no real information on patients’ level of involvement in treatment decision making, because there barely was any involvement during the consultation.
DISCUSSION
The aim of this study was to investigate the relationship between SDM and patients’ trust in their oncologist. Since observed levels of SDM were low in this study, its association with trust could not be tested sufficiently. Characteristics that were associated with trust, according to the findings in this study, were bigger tumor size and the use of Adjuvant!, the prediction model that was used as a decision support tool.
Observed/Experienced SDM and Trust
The most interesting finding in this study was that SDM and trust were not associated. The observed levels of SDM were even lower in this study than in previous studies as found by Couët et al. (2013). Patients’ level of trust in their oncologist was generally high. Given that OPTION-12 scores were low and lacked variation, it was to be expected that the objective level of SDM was not associated with trust. Patients reported strong trust even if oncologists did not practice SDM. Saba et al. (2006) suggest that communication behavior, in this case SDM, does not ensure an experience of collaboration, and a positive subjective experience of partnership does not reflect full communication. Since the trust levels that were found in this study were high, it is assumed that there was a high subjective experience of partnership. Apparently, patients felt comfortable enough during the consultations to let their oncologist take the lead, but also to express themselves if they felt that was necessary. Patients may experience some freedom of choice, even if it is not objectively visible and regardless of the way their oncologist frames their options. Patients felt they could say “no”, but very often their wishes tallied with their oncologist’s advice, so they felt they had a choice but did not explicitly use/express it. This finding could match the attitude of the modern patient, who has a need for autonomy (McKinstry et al., 2006).
Concerning the lack of adequate application of SDM during the consultations: how come there was so little SDM performed in a setting where patients’ preference should be priority? Any kind of treatment has an impact on patients’ (quality of) life, however minor it may seem. It would be ethical to let patients have the last say on whether they think the benefits are worth the costs of treatment, for it is the patient who undergoes treatment, and suffers the side-effects which can cause drastic changes in their life (Drake & Deegan, 2009). Possibly oncologists assumed they knew enough about the patient, based on their clinical experience or preconceptions about the patient. As a result, physicians may have believed that they knew what patient’s preferences were and opted not to provide them with extensive, potentially confronting information. They might have done so intending to protect or guide the patient. Also, oncologists could instinctively have felt that trust was already there and that presenting patients with more options and information would only be confusing and decrease trust, which they did not want to risk, as suggested by Hillen et al. (2011).
Alternatively, oncologists may not have perceived the different treatment options to be medically equivalent. Particularly, the option of ‘doing nothing’ may not always be viewed as a medically equally viable option in comparison with other adjuvant treatment options. Although formally, the physician was expected to present them as such (Engelhardt et al., submitted), oncologists may have left out the worst option - in their eyes - of doing nothing.
Trust-predicting characteristics
was used, patients reported higher levels of trust than after consultations in which it was not used. One explanation could be that physicians used Adjuvant! more if they had a preference for treatment in any form as opposed to no treatment at all. In other words, if adjuvant therapy was expected to lead to a significant increase in survival, the physician would use Adjuvant! for convincing purposes. As a result, patients might have perceived their physicians as more convincing in stating their arguments and using Adjuvant! to strengthen their arguments. The opposite was possible as well: if the chance that treatment would be helpful was minimal, Adjuvant! could be used by the physician to illustrate this. Patients might have consequently perceived their oncologists as highly convincing, which increases trust. In both scenarios, Adjuvant! helped the oncologist to make their point. Another explanation why the use of Adjuvant! positively influenced the levels of trust, is that the use of models and statistics during the decision-making process, might make the oncologist look more capable. This could have made patients more ready or willing to trust them, since their behavior was perceived as more scientific and objective, which in the patients’ eyes made the physician more trustworthy. In line with this explanation, earlier research has shown that patients trust a physician more, when he or she emphasizes his or her expertise (Hillen et al., 2014).
Another predictor of trust turned out to be size of the tumor: the larger the size, the more trust patients have in their oncologist. This might be caused by the greater severity of the illness, making a patient more dependent on their oncologist, who could provide any kind of help or care in their desperate situation, and hence increasing their trust automatically out of necessity. Previous research indeed suggests that high vulnerability creates the necessity to trust the physician (Hillen et al., 2012). However, this does not tally with the finding that neither tumor stage nor tumor positive lymph nodes were found to be predictors for trust, whereas both are indicators of how sick, and thus how vulnerable, patients are.
Limitations and suggestions
This study had several limitations. An important limitation was that assessing all the different ways trust can be established, is difficult, since a lot of the established trust is not only based on the words that the physicians use. It is also attained through their non-verbal contact, their tone and their
meta-communication skills (Saba et al., 2006): things that were not rated or seen in the audio-recorded consultations.
Another limitation was that the OPTION-12, the SDM instrument used in this study, only rates the physician’s actions to involve patients in the decision making process and does not take the interaction between patient and physician into account. Since the patient’s actions and reactions were not assessed, it is difficult to see how the patient’s input influenced the consultation and the physician’s attempt to apply SDM. This could be another reason why an association between SDM and trust was not found. Future research could focus more on all of the different aspects of the consultation, like for instance the non-verbal communication, meta-communication and patient’s reactions, instead of only the
oncologist’s actions. There are different instruments available to rate SDM differently, like the SDM-Q-9 and SDM-Q-Doc, which are self report instruments for both physician and patient to assess SDM (Rodenburg-Vandenbussche et al., 2015) or the Dyadic OPTION, which focuses on perceived SDM by physician and patient and thus more on the dyadic nature of the medical encounter (Melbourne et al., 2011).
Additionally, it is possible that some of the information during the consultation was not registered: the oncologist provided additional information before or after the recording ended, which most likely were not moments that had a lot of SDM, but they might have built trust without it being heard. This could also have compromised the scores on several of the OPTION-12 items, for example item 9, which checks whether the physician asks if the patient has any questions left.
In this study, potentially relevant variables which could be identified as predictors for trust, like for instance socio-economical status or ethnicity, were not available. Future studies should anticipate on this limitation and collect more information on the patients, their disease, their physicians, their hospitals and so on, to build larger and more adequate prediction models.
Despite SDM not being objectively visible in this study, it is highly recommended to examine it more closely, preferably with an experimental design. Extensive courses could be given to physicians on how to adequately apply SDM in their consultations. Their consultations should be compared to
consultations of physicians who did not have any extra courses on how to apply SDM, to see whether there is a difference in trust levels between the patients of these two groups. Another suggestion, to examine the durability of trust, is using a longitudinal design. Trust should be assessed before, during and after treatment to see whether trust levels are affected by the outcome of the chosen treatment. These designs could be combined as well.
SDM remains a relevant subject, being considered an ethically just approach in medical decision making – some might argue that SDM is the golden standard within modern physician-patient communication (Hauser, Koerfer, Kuhr, Albus, Herzig & Matthes, 2015). It needs to be assessed more thoroughly in the future, since its adequate implementation might influence trust, i.e., increase or decrease it. Both
scenarios would have consequences and for the educational purposes of physicians, more substantiated conclusions are needed, so they know whether and how to implement SDM in their relationship with their patients.
Conclusions
SDM was not adequately implemented in the vast majority of consultations. Nevertheless, patients’ trust in their oncologist was high. This suggests that the application of SDM during the consultation is not associated with the level of trust. However, the only true conclusion that can be drawn, is that there was too little SDM implemented to make any statements about a potential association with trust. It is,
however, premature to conclude what the implications of this finding could be. Future research needs to assess what higher levels of SDM during the consultation would mean for the level of trust.
REFERENCES
Abel, G. A., Roland, M., Campbell, J. L., Shay, L. A., & Lafata, J. E. (2015). Where is the evidence? A systematic review of shared decision making and patient outcomes. Med Decis Making, 35(1), 114-31.
Akaike, H. (1971). Information theory and an extension of the maximum likelihood principle. Paper presented at: 2nd International Symposium on Information Theory, Tsahkadsor, Armenia.
Bursac, Z., Gauss, C. H., Williams, D. K., & Hosmer, D. W. (2008). Purposeful selection of variables in logistic regression. Source Code Biol Med, 16, 3:17.
Butterworth, J. E., & Campbell, J. L. (2014). Older patients and their GPs: shared decision making in enhancing trust. The British journal of general practice : the journal of the Royal College of General Practitioners, 64(628), pp.e709-18.
Couët, N., Desroches, S., Robitaille, H., Vaillancourt, H., Leblanc, A., Turcotte, S., et al. (2013). Assessments of the extent to which health-care providers involve patients in decision making: a systematic review of studies using the OPTION instrument. Health Expectations, 18(4), 542-61.
Croker, J. E., Swancutt, D. R., Roberts, M. J., Abel, G. A., Roland, M., & Campbell, J. L. (2013). Factors affecting patients’ trust and confidence in GPs: evidence from the English national GP Patient Survey. BMJ Open, 3(5), e002762.
Degner, L. F., & Sloan, J. A. (1992). Decision making during serious illness: what role do patients really want to play? J Clin Epidemiol 45(9):941-950.
Drake, R., & Deegan, P. (2009). Shared Decision Making Is an Ethical Imperative. Psychiatric Services, 60(8), 1007.
Elwyn, G., Hutchings, H., Edwards, A., Rapport, F., Wensing, M., Cheung, W. Y., et al. (2005). The OPTION scale: measuring the extent that clinicians involve patients in decision-making tasks. Health Expect. Mar;8(1):34-42.
Elwyn, G., Tsulukidze, M., Edwards, A., Légaré F., Newcombe, R. (2013). Using a 'talk' model of shared decision making to propose an observation-based measure: Observer OPTION 5 Item. Patient Educ Couns, 93(2):265-71.
Engelhardt, E. G., Pieterse, A. H., Van Duijn-Bakker, N., Quarles van Ufford-Mannesse, P., Sleeboom, H., Portielje, et al. (submitted). Information provision about the benefits and side-effects of adjuvant systemic therapy for breast cancer in clinical practice: does the use of Adjuvant! facilitate communication?
Faul, F., Erdfelder, E., Buchner, A., & Lang, A.-G. (2009). Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behavior Research Methods, 41, 1149-1160.
Fuertes, J., Mislowack, A., Bennett, J., Paul, L., Gilbert, T. C., Fontan, G., et al. (2007). The physician patient working alliance. Patient Educ Counsel, 66(1), 29–36.
Hauser, K., Koerfer, A., Kuhr, K., Albus, C., Herzig, S., & Matthes, J. (2015). Outcome-Relevant Effects of Shared Decision Making: A Systematic Review. Dtsch Arztebl Int Oct 2; 112(40): 665–671. doi: 10.3238/arztebl.2015.0665.
Hillen, M. A., De Haes, J. C .J. M., & Smets, E. M. A. (2011). Cancer patients’ trust in their oncologist – a review. Psycho-oncology, 20(3), 227-241.
Hillen, M. A., De Haes, H. C. J. M., Stalpers, L. J. A., Klinkenbijl, J. H. G., Eddes, E. H., Butow, P. N., et al. (2014). How can communication by oncologists enhance patients’ trust? An experimental study. Annals of Oncology, Vol. 25(4), pp.896-901.
Hillen, M. A., Koning, C. C. E., Wilmink, J. W., Klinkenbijl, J. H. G., Eddes, E. H., Kallimanis-King, B. L., et al. (2012). Assessing cancer patients’ trust in their oncologist: development and validation of the Trust in Oncologist Scale (TiOS). Supportive Care in Cancer; 20(8), 1787-1795.
Hillen, M. A., Onderwater, A. T., Van Zwieten, M. C. B., De Haes, J. C. J. M., & Smets, E. M. A. (2012). Disentangling cancer patients’ trust in their oncologist: a qualitative study. Psycho-oncology, 21(4), 392-399.
Hosmer, D. W. (2013). Applied Logistic Regression. New York: John Wiley & Sons, Inc. pp. 91-142. McKinstry, B., Ashcroft, R. E., Car, J., Freeman, G. K., & Sheikh, A. (2006). Interventions for improving patients’ trust in doctors and groups of doctors. Cochrane Database Syst Rev, 3:CD004134.
Mechanic, D. (1996). Changing medical organization and the erosion of trust. Milbank Quart, 74(2), 171–189.
Mechanic, D., & Meyer, S. (2000). Concepts of trust among patients with serious illness. Soc Sci Med, 51(5), 657–668.
Melbourne, E., Roberts, S., Durand, M., Newcombe, R., Légaré, F., & Elwyn, G. (2011). Dyadic OPTION: Measuring perceptions of shared decision-making in practice. Patient Education and Counseling, 83(1), 55-57, ISSN 0738-3991, https://doi.org/10.1016/j.pec.2010.04.019.
Ravdin, P. M., Siminoff, L. A., Davis, G. J., Mercer, M.B., Hewlett, J., Gerson, N., et al. (2001). Computer program to assist in making decisions about adjuvant therapy for women with early breast cancer. JClinOncol.19(4):980-991. http://jco.ascopubs.org/content/19/4/980.long.
Ridd, M., Shaw, A., Lewis, G., & Salisbury, C. (2009). The patient doctor relationship: a synthesis of the qualitative literature on patients’ perspectives. Br J Med Prac, 59(561), 116–133.
Rodenburg-Vandenbussche, S., Pieterse, A. H., Kroonenberg, P. M., Scholl, I., Weijden, T. T., Van der Luyten, G. P. M., et al. (2015). Dutch Translation and Psychometric Testing of the 9-Item Shared Decision Making Questionnaire (SDM-Q-9) and Shared Decision Making Questionnaire-Physician Version (SDM-Q-Doc) in Primary and Secondary Care. PLoS One, 10(7), E0132158-6203.
Saba, G., Wong, S., Schillinger, D., Fernandez, A., Somkin, C., Wilson, C., et al. (2006). Shared decision making and the experience of partnership in primary care. Annals Of Family Medicine, 4(1), 54-62.
Skirbekk, H. (2009). Negotiated or taken-for-granted trust? Explicit and implicit interpretations of trust in a medical setting. Med Health Care Philos, 12, 3–7.
Stiggelbout, A. M., Pieterse, A. H., De Haes, J. C. (2015). Shared decision making: Concepts, evidence, and practice. Patient Educ Couns. 98(10):1172-9.
Stiggelbout, A. M., Van der Weijden, T., De Wit, M. P., Frosch, D., Légaré, F., Montori, V. M., et al. (2012). Shared decision making: really putting patients at the centre of healthcare. BMJ, 27, 344-356.
World Health Organization (1994, November). A declaration on the promotion of patients’ rights in Europe. Retrieved on June 1 2017 from http://www.who.int/genomics/public/eu_declaration1994.pdf.
APPENDIX
Table 1: Items of the Observing Patient Involvement In Decision-making rating scale (OPTION-12). Elwyn et al. (2005).
1. The clinician draws attention to an identified problem as one that requires a decision making process.
2. The clinician states that there is more than one way to deal with the identified problem (‘equipoise’).
3. The clinician assesses the patient’s preferred approach to receiving information to assist decision making (e.g. discussion, reading printed material, assessing graphical data, using videotapes or other media).
4. The clinician lists ‘options’, which can include the choice of ‘no action’.
5. The clinician explains the pros and cons of options to the patient (taking ‘no action’ is an option).
6. The clinician explores the patient’s expectations (or ideas) about how the problem(s) are to be managed.
7. The clinician explores the patient’s concerns (fears) about how problem(s) are to be managed. 8. The clinician checks that the patient has understood the information.
9. The clinician offers the patient explicit opportunities to ask questions during the decision making process.
10. The clinician elicits the patient’s preferred level of involvement in decision-making. 11. The clinician indicates the need for a decision making (or deferring) stage. 12. The clinician indicates the need to review the decision (or deferment).
Table 2: Items of the Trust in Oncologist Scale (TiOS). Hillen et al. (2012). 1. Your doctor is very careful and precise
2. Your doctor is totally honest in telling you about all the different treatment options available for your condition
3. Your doctor always gives you honest information about your prospects 4. Your doctor strongly cares about your health
5. Your doctor always tells you everything you want to know about your illness 6. You think your doctor can handle any medical situation, even a very serious one 7. Your doctor always takes his/her time with you.
8. Your doctor explains everything so that you can consent to medical decisions 9. Sometimes you worry that your doctor's medical decisions are wrong*
10. Your doctor only thinks about what is best for you
11. Sometimes your doctor does not pay full attention to what you are trying to tell him/her* 12. Your doctor would always tell you the truth about your health, even if there was bad news 13. You have doubts whether your doctor really cares about you as a person*
14. Your doctor listens with care and concern to all the problems you have 15. Your doctor will do whatever it takes to get you all the care you need 16. Your doctor is available for you whenever you need him/her
17. You have no worries about putting your life in your doctor's hands 18. All in all, you have complete trust in your doctor
Note: Items with * are reverse-scored.
Table 3: Variables included in the multivariable analyses.
Patient Characteristics (all women)
1. Age (in years)
2. Education (high/ intermediate/low) 3. Children (yes/no)
4. Marital status (single/in a relationship)
5. Preferred level of involvement in decision-making (patient-driven/shared/oncologist-driven)
Hospital/Oncologist Characteristics
6. Hospital affiliation (peripheral/academic) 7. Gender Oncologist (male/female)
8. Use of Adjuvant! Online software during consultation (yes/no) 9. Duration consultation (hours:minutes:seconds)
Disease Characteristics 10. TNM Stage of disease: Stage I Stage II/III 11. Size of tumor Small (0-20mm) Big (>21mm)
Table 4: Descriptives of Sample Characteristics.
Demographics N = 101
Age in years, mean (± SD) 60.6 (11)
Education High Intermediate Low 29 (29) 51 (50) 21 (21) Marital status Single In a relationship 24 (24) 77 (76) Has children 84 (83)
Preferred level of involvement in decision-making Patient-driven Shared Oncologist driven 39 (39) 41 (41) 21 (21) Consultation characteristics
Consultations with University hospital oncologists 20 (20) Consultations with use of Adjuvant! Online software 60 (59)
Consultation duration, mean (± SD) 00:28:03 (00:12:29)
Scores
TiOS score, mean (± SD) 4.05 (0.56)
OPTION-12 sum score, mean (± SD) 15.5 (11.6)
Disease characteristics Tumor size (mm) Small (0 – 20) Big (≥ 21) 57 (56) 42 (41) Tumor Stage I II/III 42 (42) 56 (56)
Tumor Positive Lymph Nodes present 32 (32)
Note: Data are shown as n (%) unless otherwise specified. Abbreviations: CPS: Control Preferences Scale; N: number; SD:
standard deviation; TiOS: Trust in Oncologist Scale.
Table 5: Variables associated with increased trust.
B SE Beta P
Constant 68.312 1.756 <.001
Tumor Size, > 20mm 4.265 1.993 .210 .035
Figure 1. Patients’ answers as to whom they thought had made the final decision concerning their treatment.
Note: Only the patients with whom a specific treatment was discussed that decided whether to undergo treatment and who
gave a clear answer that could be categorized are included in this figure: n =71 for chemotherapy and n = 78 for hormonal therapy.
66% 26%
8%
Chemotherapy
Patient decided Oncologist decided Decided together
46% 45%
9%
Hormonal Therapy
Patient decided Oncologist decided Decided together
