Barriers in Advance Care Planning and Goals of Care Discussion With Oncology Patients: The Perspective of Brazilian Medical Residents

Background. Advance care planning (ACP) and goals of care discussions involve the exploration of what is most important to a person to prepare for health-care decision-making. Despite their well-established benets, they are still not frequently performed in clinical oncology practice. The awareness of physicians’ barriers enables the development of tailored interventions to improve professional practice. Considering the scarcity of studies about the diculties of medical residents in discussing goals of care, this study aims to determine, from the perspective of medical residents, the barriers to the discussion goals of care and ACP with patients with cancer. Methods. A previously published survey (Decide-Oncology) to assess barriers to goals of care discussion among health care providers was adapted to Portuguese language and distributed to Brazilian medical residents from three hospitals in Brazil in a cross-sectional pilot study. Residents were asked to rank the importance of various barriers to discuss goals of care (ranging from 1-extremely unimportant to 7-extremely important). Results: Twenty-nine residents answered the questionnaire (30.9%). The most reported barriers were related to patients and their families' diculty in understanding and accepting the diagnostic and the prognosis, patients’ desire to receive full active treatment and the lack of an advance directive. Furthermore, the physician and external factors such as lack of training and lack of time to have these conversations were also very important barriers. The identication of the key barriers that limit the discussion of ACP and early palliative care referrals can certainly help to prioritize the next steps for future studies aimed at improving ACP and goals of care discussions.


Introduction
Advance care planning (ACP) and goals of care discussions involve the exploration of a person's values, beliefs, and what is most important to each person: to ensure concordance between the clinical care received by a patient and the care he or she has wished for. The context for each differs, however, as ACP conversations focus on preparing for future health-care decisions, whereas goals of care discussions focus on current health-care decisions. Thus, goals of care discussions are also an important part of the advance care planning process. 1,2 One of the principles of good care is respect and receptivity to patients' wishes and values; thus, it is important to understand patients' perspectives about cancer treatment. 3 High-intensity life support is often provided even when patients or their families may prefer to focus on comfort care, in spite of evidence suggesting that use of unwanted technology is associated with negative outcomes, such as decreased quality of life and low satisfaction with end-of-life care 4 . Robust evidence shows that ACP is associated with a wide variety of bene ts, such as: less moral distress by health care professionals, higher rates of advanced directives (AD) registrations, reduced hospitalization, intensive and futile treatment at the end-of-life, greater probability of the patient dying in the chosen place, greater satisfaction with the quality of care, and less risk of stress, anxiety and depression in family members during bereavement. 1,5 All in all, it improves the consistency of care with patients' values and goals in various patient populations, including oncology patients. 6 Notwithstanding the importance of ACP, many health-care providers are still reluctant to engage patients in this discussion. Some barriers include concerns about causing distress for a person or decreasing hope for life-sustaining outcomes, personal discomfort with death and dying, lack of experience in discussing end-of-life issues and limited education and training in ACP 1 . By the same token, implementing interventions to improve communication and decision making about goals of care and ACP in oncology settings, requires an understanding of the perspectives of patients, families and clinicians. Likewise, awareness of barriers enables the development of tailored interventions that are more likely to improve professional practice and training programs 7,8 . Considering the scarcity of studies about the di culties of medical residents on discussing ACP and about their training on this topic, this study aims to determine, from the perspective of medical residents, the barriers to the discussion goals of care and ACP with patients with cancer.

Aim
To identify barriers to goals of care and advance care planning discussions in patients with advanced cancer from the perspective of medical residents.

Study Questionnaire
This cross-sectional pilot study with Brazilian medical residents used the semi-structured questionnaire "Decide Oncology", a Canadian questionnaire developed and adapted to the purpose of identifying and evaluating the di culties that impact on the advance care planning of cancer patients, according to the perception of doctors who assist patients in oncology settings. Resident physicians were also included in its validation; therefore, it was the chosen tool for this study 6 . Statements about doctors' di culties focused on three pillars: (1) barriers on the discussion about goals of care; (2) barriers on the approach of interrupting active cancer therapies (for example: chemotherapy, radiation therapy); and (3) barriers of referring to palliative care teams. These di culties were further strati ed into barriers related to patients, doctors and/or related to the health system or external factors. It is a Likert questionnaire where each session has statements about the di culties in discussing the advance care planning, classi ed from 1 to 7 points, where 1 is the extremely unimportant di culty; 2, very unimportant; 3, little unimportant; 4,neither important nor unimportant; 5, little important; 6,very important; and 7, extremely important 3 .
The questionnaire also included questions about sociodemographic characteristics (age, gender, ethnicity, religion, year of graduation and medical fellowship and specialty), and questions about the perception of the importance and the degree of formal training on ACP and discussing goals of care.
The adaptation and translation of the instrument Decide-Oncology was conducted. The direct and reverse translation, from English into Portuguese (Brazil), by four translators was done followed by the evaluation of six health professionals who are part of palliative care teams (three doctors, a nurse, a social worker and a psychologist), invited as judges to analyze the clarity of the language, content validity and the translation's adequacy. For this, the Content Validity Coe cient (CVC) was used, as suggested by Hernández-Nieto (2002). At the end of the judgment, the criterion of CVC > 0.80 was used to consider whether the items have been semantically adjusted 9 .
The questionnaire was adapted into an electronic version (Google forms) and sent via email to medical residents from three Brazilian reference oncology centers of university hospitals (Belém-Pará, Recife and São Paulo) and ten residents were invited to answer the paper questionnaire during a workshop about breaking bad news.
The study was submitted and approved by the Ethics Board of the João de Barros Barreto University Hospital (Belém-Pará-Brasil). All participants provided written informed consent.

Study population
Resident physicians who work or who have worked in oncology and palliative medicine services (e.g. oncology, internal medicine, geriatrics and hematology), from three University Hospitals were contacted. Out of ninety-four resident physicians referred, 29 responded. Both paper and electronic surveys were used to increase response rate. It must be noted that possible duplications were excluded.

Statistical Analysis
The results of paper and electronic surveys were pooled and compiled using descriptive statistics, including median and interquartile interval [1st Q; 3rd Q] for quantitative variables and counts and proportions for categorical variables. Survey responses were presented by median scores. Distribution of scores were compared between groups of residence years using Mann-Whitney's non-parametric test. Data analysis was performed using SPSS® Statistics (version 26.0; SPSS Inc., Chicago, IL, USA). In all tests, values of p < 0.05 were considered signi cant.

Participants
The survey was distributed to a total of 94 medical residents (10 paper, 84 electronic forms) between December 2020 and March 2021. A response rate of 30.9% (100% answered the paper questionnaire and 22.6% the electronic version) was achieved. Respondent demographics are outlined in Table 1. Hematology 6 (20.7) Palliative medicine 2 (6.9) Clinical Oncology 2 (6.9) a One participant did not answer this questio Barriers to discuss ACP and goals of care Most of the residents described the patient and family barriers as the most important, such as "patients' di culty in understanding their diagnostic and on accepting their prognosis", "patients' desire to receive full active treatment", "the lack of an advance directive", "family members' di culty in understanding their loved one's prognosis". About the physician's barriers, "lack of training" was the most important barrier described. "Lack of time to have conversation" and "Lack of availability of substitute decision maker(s)" were two of the barriers described as very important. The importance given to the "lack of time to have conversations about goals of care" was signi cantly different between the two groups of years of residency(p = 0.034), being a more important barrier among those with more years of residency (median scores of 4 and 6 for groups of up to 2 years and more than 2 years of residency, respectively, Table 2) Barriers to discuss the discontinuation of cancer-directed therapies The barriers with higher scores were the patient and family barriers, which were also most frequently described to the discontinuation of cancer-directed therapies. Additionally, patients' expectations regarding the bene ts of therapy. Also, patients' and families' poor appreciation of prognosis or denial of likely survival duration, as well as patients' in ated expectation of the bene ts from further cancerdirected therapy. The most described physicians' barriers were the discordance with other specialists in estimating prognosis/length of survival, followed by di culty on estimating patient prognosis/length of survival, uncertainty of the bene ts of further active cancer therapy, and patient age (whereby had more di culty to suspend active treatment in younger patients). External barriers were ranked lower than patient/family and physician-level barriers. (Fig. 1B). Therefore, comparing the two groups of residents, "patients' expectations regarding the bene ts of therapy" and "patient age" were the barriers that presented signi cant different scores and higher importance in the group with more than two years of residency (p < 0.001 and p = 0.025, respectively, Table 2).
Barriers to early palliative care referrals The external barriers such as lack of access to palliative care services in the community and in the hospitals, lack of multidisciplinary team (social workers, nurse practitioners, etc) to aid in patient support/referral process to palliative care, and the patient or family refusal of referral, were the barriers most described to early palliative care referrals (Fig. 1).
The willingness to discuss ACP and goals of care Most of the respondents were very or extremely willing to initiate (82.8%) and lead (86.2%) ACP and goals of care discussions.
Physician's perception about the ability to discuss ACP and goals of care and the quality of training on communication.
Most of the residents' self-reported average skill in having goals of care discussions with patients and families (44.8%); while 89.6% ranked their priority for learning these skills as high (as 4 or 5 out of 5). By contrast, 93.1% had formal training regarding communication and goals of care. Besides, 77.7% considered their quality of training from moderate to extremely high (Table 3). Sixteen participants (56%) suggested continuous education in palliative care in communication skills and on advance care planning as tools for improving goals of care conversations, like this one: "Training about breaking bad news skills, palliative care and death during graduation and medical residency. Little is taught about it". Thirty seven percent of those who suggested how to improve this communication cited the importance of an interdisciplinary team aligned in the discussion of care: "Constant training of the medical and multidisciplinary team, as well as regular meetings with a multidisciplinary and palliative care team, also involving patients and family members". Some residents (18%) also described that a good physician-patient relationship could help to improve communication and the decision-making process about goals of care.

Discussion
The frequent barriers to the discussion of goals of care and ACP present in the literature included lack of time, discomfort with di cult discussions, fear of affecting patients' hope and emotional coping and lack of training in communication strategies about end-of-life 5,10,11,12 . In the present study, the "Decide-Oncology" questionnaire 4 was used to evaluate the barriers on discussing goals of care. It classi es the barriers between: those of physicians, those of patients and families, and those of external barriers. Regarding the di culties to discuss goals of care, most were related to patients' and family members' factors, such as the di culty for patients to understand and accept their prognosis. These ndings corroborate those of the Decide-study 7 and Decide-Oncology study 4 , in which health-care professionals on general medical wards and physicians in oncology ambulatory identi ed patient and family related factors as the most important barriers to goals of care discussion. Some external factors such as the lack of time for these discussions was also considered a signi cant barrier in the present study, in agreement with the aforementioned authors 4,7 .
However, the physicians' perception about the patient's di culty to understand and accept his/her prognosis may re ect, at least in part, on the doctors' di culty in predicting prognosis -a well-reported barrier in this study -and even with the well-established prognosis, they may have di culties on communicating this prognosis in an objective and clear way 4,7 . It is important to consider that this study involved exclusively medical residents -who spend at least one month at oncology and/or palliative care services-unlike the DECIDE oncology study 4 , which included mostly oncologists, possibly with more experience than medical residents.
When comparing groups of residents by years of medical residency, 3rd and 4th year residents (generally on oncology, geriatric, hematology or palliative medicine residency) considered having more di culties to discuss about goals of care related to the patients' expectations regarding the bene ts of therapy, and  15,17 . In contrast, patients who understood that the role of chemotherapy was to reduce disease burden and control symptoms were more likely referred to appropriate hospice services and receive less aggressive care at the end-of-life 18 . Efforts to optimize communication are essential to effectively address gaps in illness understanding and moderate expectations for bene t, while preserving hope, empathy, and the therapeutic relationship between patients and their healthcare team 19 .
Granek (2013) explored oncologists' views on end-of-life discussions and revealed various physicianrelated barriers to ACP, including personal discomfort with death and dying, reluctance to communicate painful information, perception of other physicians' failure, focusing too much on a cure or treatment, along with a lack of experience with breaking bad news, especially in end-of-life care and lack of mentorship on training end-of-life communication strategies 12 . Regarding the discontinuation of directed cancer therapies, this study also found the discordance with other specialists in estimating prognosis/length of survival, followed by di culty on estimating patient prognosis and uncertainty of the bene ts of further active cancer therapy as important barriers. It is known that earlier follow-up by palliative care team could improve the communication with patients with incurable disease in helping them to prepare for the end of life, through shared decision-making among patients, their families, and healthcare providers 20 .
There is evidence that the use of chemotherapy near the end of life is not related to its likelihood of providing bene t. So, the suspension of aggressive treatment before the end of life is an indicator of good quality care, especially in the last two weeks of life 21 .This indicator is often di cult to capture but re ects the notion that optimal palliative care in the nal weeks of life generally includes withdrawal of cancer aggressive and directed therapies, including withholding aggressive resuscitation measures 22 .
This concept may be more di cult for patients to comprehend, as the mistaken notion of ongoing cancer treatment is often connected to maintaining good care and hope 4 . In the present study, medical residents indicated lack of access to palliative care services in the community and at hospitals, and lack of multidisciplinary teams (social workers, nurse practitioners, etc.) hindered patient support or referral process to palliative care, as well as the patient or family refusal of referral to palliative care, as opposed to the Decide-Oncology study 4  In the present study, most of the residents self-reported an average skill level to discuss goals of care with their patients and also ranked it as high priority to learn those skills, although most of them already had a formal training regarding communication and goals of care discussion, which they classi ed as a moderately high-quality training. It is probably due to the fact that most of the residents are already at the third or fourth year of residency, in oncology-hematology or palliative care services at university hospitals, and deal with patients with advanced diseases frequently in their clinical practice.
Different tools may be more or less effective for clinicians with varying expertise and competency in end- Barriers related to patients and family members were the most reported in the present study, as well as in previous studies involving only health care professionals 4,7 .
In Brazil, a competency framework of palliative medicine for geriatricians' residents was developed to guide the teaching of essential and desired skills of palliative care in geriatrics settings 34 . It includes communication's skills such as: to explore the elderly patient and family's understanding of illnesses, concerns, goals and values and planning treatment that ts these priorities (advance care planning); to perform an effective patient-centered communication on breaking bad news or delivering prognostic information, in addition to adequately communicating to the patient on how to perform advance directives, based on bioethical and legal principles 34 . In other areas such as oncology, a competency framework would certainly help to standardize the skills and competencies of resident physicians. In our study, a great majority of residents described continuing education in palliative care and communication skills training as suggestions for improving shared decision-making about goals of care.
Given the lack of studies that address barriers to discussing goals of care, especially in the Brazilian reality, evaluating the di culties from the perspectives of medical residents and their relative importance to discussing this topic are the strengths of this study. This study is the rst step for the adaptation and validation of the Decide-oncology questionnaire to the Brazilian-Portuguese language. Nonetheless some limitations included a small sample, because it involved only medical residents from three university hospitals; due to this, it is possible that our ndings may not be representative of all eligible respondents. Furthermore, a positive response bias among those who did participate is possible, with potential for higher response rates among those most engaged in palliative and end-of-life care.

Conclusion
This study evaluated the medical residents' di culties in discussing goals of care and ACP with oncology patients. The most important reported barriers were related to patients and their families, such as their di culty in understanding and accepting the diagnostic and the prognosis. The study also described patients' desire to receive full active treatment and the lack of an advance directive. Similarly, the physician and also external factors such as lack of training and lack of time to have these conversations were equally very important barriers. In addition, the importance and willingness to initiate and lead ACP and goals of care discussions were demonstrated by most residents in this study, who consequently indicated the need for training in this area as a priority. Not to mention that there are various interventions to prepare patients and families for ACP and end-of-life communications, strategies to overcome barriers for healthcare providers and practical tips to initiate these di cult conversations 1,8,23,33 . Moreover, the identi cation of the key barriers that limit the discussion of ACP and early palliative care referrals could help to prioritize the next steps for future studies aimed at improving advanced care planning and goals of care discussions 7 . All things considered, communication skills training and tools that enhance medical residents and clinicians' ability to listen with empathy, to strengthen doctor-patient relationship and to discuss clearly prognosis, will de nitely help clinicians to better support patients and families through shared decision-making based on the patient's values and experiences.

Declarations
Ethics approval and consent to participate Ethical approval was waived by the Ethics Board of the João de Barros Barreto University Hospital (Belém-Pará-Brasil). Informed consent was obtained from all subjects. All methods were carried out in accordance with relevant guidelines and regulations.

Consent for publication
Not applicable.

Figure 1
Barriers of patient, physician, and external factors to initiating goals of care discussions (A), to interrupt cancer-directed therapies (B), and to refer to palliative care (C) (median scores).