The objectives of this study were to describe DC patients' preferences for autonomy and involvement in therapeutic decision-making and to identify variables associated with these preferences.
Control Preference Scale
In this study, the proportion of patients who preferred a passive or a collaborative role in therapeutic decision-making were similar (49.1% and 43.0%, respectively). The proportion of patients who preferred an active role was low (7.9%). The distribution of roles differs greatly from one study to another. For instance, Elkin et al. conducted a study in metastatic colorectal cancer patients over 70 years old and found that 52% of patients preferred a passive role, 23% of patients a collaborative role and 25% of patients an active role [22]. In contrast, a study conducted by Bruera et al. in patients suffering from mixed types of cancer reported a high preference for the collaborative role (63%) [23]. These discrepancies could be related to the characteristics of the population studies, which widely differ. For instance, the average age of the Elkin et al. population was higher than the mean age of the population, which might explain the lower preference for the collaborative role. Furthermore, it has been found in several studies that women are generally collaborative [24,25], confirmed in the female dominated population of Bruera et al. Noteworthy, differences in care systems according to countries that allow more or less room for patient preferences could also be associated with these discrepancies. Turning now to the variables significantly associated to the preferences in our study, occupational status was related to the desire for autonomy. The proportion of patients reporting a passive role in decision-making was higher among retired DC patients than among employed DC patients. Perhaps less involvement in occupational life and less responsibility would favor increased passivity in decision-making. The variable "time since diagnosis" was also related to CPS status. Patients who reported a preference for collaborative decision-making had their diagnosis for a longer time. Increased awareness about patient autonomy could better support the expression of patient preferences and their identification by the oncologist could support shared decision-making.
Autonomy Preference Index DM score
The mean DM score in our study was 51.5. Another study in primary care found scores between 36.6 and 50.6 depending on the motives for consultation. Lower scores were associated with increased severity of the disease and older age [26]. Considering digestive cancers as severe diseases, the average score of autonomy in our study could be interpreted as high, which reflects a strong desire of patients to be involved in their treatment. None of the variables tested in this study were significantly associated with the DM score.
The desire for information characterized by the API IS score was neither correlated with the API DM score nor with the CPS status. We expected that the more patients would like to be involved in decision-making, the more information they would like to have about the treatment and the disease. A systematic review identified several factors associated with the need for health information, such as gender, age, education level, time since diagnosis, course of the disease and psychological QoL, but no associations were found with the desire for autonomy in decision-making [27].
Anxiety, depression and health-related QoL were not significantly correlated with the desire for autonomy and involvement in decision-making in our study. We hypothesized that non-anxious and non-depressed patients would like to be more involved in decision-making. Anxiety is characterized by ruminations that can interfere with decision-making [28] and depression is characterized by a state of sadness that leads to withdrawal and low commitment in actions [29].
The originality of this descriptive study was to identify the DC patients' preferences in relation with their autonomy and involvement in therapeutic decision-making. Our preliminary study can help to determine variables that have to be taken into account in further in-depth investigations relevant for clinical practice. Previous studies showed that patients often do not have the decision-making roles they would like to have [30,31] or that physicians' perceptions of patients' preferences can be wrong [22]. In parallel, other studies have shown that the more a decision is collaborative, the more a patient is satisfied with the choice of treatment and the decision-making process [32,33]. In the current context of changing medical culture, shared decision-making is often considered as the ideal model for the physician-patient relationship. Medical culture is moving from medical paternalism [34], which means that physicians make decisions with little room for patient preferences, to a system that includes patient autonomy and informed participation in medical decisions [35]. Medical shared decision-making honors the knowledge and the skills of the physician on one side and the rights of patients to have their preferences informed by decisions on the other side. However, shared decision-making is not wanted by all patients, especially those who prefer to have a passive role. Moreover, the experience of a specific disease is likely to affect patients' preferences and their desire for involved decision-making. This desire can evolve over time and change with the stage of disease [24].
In France, the 2002 law strengthens the patients’ right to make decisions and gives them legislative tools to promote their autonomy. This autonomy requires an understanding of the situation and of the therapeutic possibilities in order to make a choice. In this context, the law stresses the need to inform and on the quality of the delivered information. The oncologist must explain to the patient the situation and his/her reasoning from the assessment, i.e. treatment options, benefit/risk balance of each option with the therapeutic aim. This information is necessary to promote patient autonomy in decision-making. It has to be accurate and individually adapted to the patient to facilitate cognitive and emotional processing. The opinion of an oncologist can differ from the opinion of a patient if a common goal was not previously defined. This step is particularly significant as they potentially rely on different values. Regarding principles of biomedical ethics [1], the promotion of patient autonomy is linked to the principle of beneficence: the moral imperative of seeking the greatest benefit for the patient. This autonomy principle should enable physicians to seek out skills to explore and highlight patients' preferences. On the other hand, involvement in therapeutic decision-making has to rely on patient desire, since involving a patient who does not want to make healthcare decisions could have the reverse effect and decrease patient autonomy, which would be ethically harmful. It would dishonor the principles of beneficence and non-maleficence.
Limitations and perspectives
There is a possible inclusion bias since patients who are interested in shared decision-making might have been over-represented. A second limitation was the heterogeneity of time between the last therapeutic decision and the completion of questionnaires. As the experience of the disease may affect patients' preferences for autonomy [24], it would be beneficial to include patients at time of cancer diagnosis and to follow them over their course of treatment and recovery. This may enable an increased insight into factors associated with preferences and to study change in preferences over time. Finally, in the literature, level of education was positively correlated with the preference for an active or a collaborative role in decision-making [25,36,37]. It could be interesting to conduct an ancillary study using semi-structured interviews to explore the experience of patients regarding their desire for autonomy and involvement. In addition, complementary qualitative data could help to explain quantitative results and to identify relevant variables to investigate in future studies.