Gaps in patient-physician communication at the time of malignant bowel obstruction from recurrent gynecologic cancer: a qualitative study

We sought to investigate the patient and physician approaches to malignant bowel obstruction (MBO) due to recurrent gynecologic cancer by (1) comparing patient and physician expectations and priorities during a new MBO diagnosis, and (2) highlighting factors that facilitate patient-doctor communication. Patients were interviewed about their experience during an admission for MBO, and physicians were interviewed about their general approach towards MBO. Interviews were analyzed for themes using QDAMiner qualitative analysis software. The analysis utilized the framework analysis and used both predetermined themes and those that emerged from the data. We interviewed 14 patients admitted with MBO from recurrent gynecologic cancer and 15 gynecologic oncologists. We found differences between patients and physicians regarding plans for next chemotherapy treatments, foremost priorities, communication styles, and need for end-of-life discussions. Both patients and physicians felt that patient-physician communication was improved in situations of trust, understanding patient preferences, corroboration of information, and increased time spent with patients during and before the MBO. Gaps in patient-physician communication could be targeted to improve the patient experience and physician counseling during a difficult diagnosis. Our findings emphasize a need for patient-physician discussions to focus on expectations for future cancer-directed treatments, support for patients at home with home health or hospice level support in line with their wishes, and acknowledgement of uncertainty while providing direct information about the MBO diagnosis.


Introduction
A malignant bowel obstruction (MBO) is a turning point in the care of women with a gynecologic cancer. From the patient perspective, MBO is associated with many distressing symptoms including nausea, vomiting, pain, and inability to eat that are difficult to control for many women who may have been previously only mildly symptomatic during initial treatment for recurrent disease. Physicians reevaluate whether to stop previously established treatment plans, including chemotherapy, until MBO resolution or indefinitely.
Literature on MBO prognosis and interventions shows a median survival after MBO ranging between 1 and 6 months [1][2][3][4][5]. The trajectory and implications of a MBO diagnosis in the setting of recurrent ovarian cancer is not clear to many women with ovarian cancer [6]. A study in the setting of end-of-life communication showed that although oncologists report discussing prognosis or palliative intent of interventions such as chemotherapy, at least one-third of patients were unable to reflect that understanding [7]. Communication strategies have focused on how to bridge such gaps. For example, qualitative studies have shown that patients prefer serious news disclosures to occur in quiet places, one-onone, with a physician who can give a comprehensive evaluation of the diagnosis, prognosis, and treatment options [7]. In prognostic discussions and transitions from curative to end-of-life care, patients have highlighted the importance of clarity of information, grieving loss, and sustaining hope [7,8].
In this qualitative study, our objective was to explore physician and patient approaches to communication and treatment management for the diagnosis of MBO due to recurrent gynecologic cancer in order to identify areas of misalignment that can be identified and targeted with communication strategies specific to the MBO experience. Improved communication could enrich the patient-centeredness of discussions at the time of MBO by keeping patient priorities at the forefront and communicating information patients need but may not be receiving. We interviewed patients during the MBO admission at two academic-affiliated hospitals and gynecologic oncologists who frequently cared for women with MBO at these and other hospitals. Our goals were to (1) compare patient and physician expectations and priorities during a new MBO diagnosis, and (2) highlight factors noted by patients and physicians to facilitate patient-doctor communication in the setting of MBO.

Methods
Patient interviews were conducted at two institutions (the University of Chicago and Northshore University Health System). The patient interviews were granted IRB approval and cancer center approval at both institutions at which the interviews were conducted, and the physician interviews were granted an IRB exemption through the University of Chicago. The patient census was followed, and those admitted for a MBO between 5/2016 and 10/2018 were flagged by the primary researcher (CH). The primary researcher was a research fellow. During her clinical years, she was involved in the care of the last three patients interviewed and worked with seven of the interviewed gynecologic oncologists. No compensation was provided for participation in the study.
Patient inclusion criteria included malignant bowel obstruction based on clinical findings and/or imaging; known recurrent; or progressive gynecologic cancer with previous treatment, age , English-speaking, and able and willing to participate in an interview. Patients considering hospice after discharge from the hospital were included.
The charts were reviewed, and after discussion with the primary team to ensure the patient was a suitable candidate, she was approached by the researcher regarding participation in the study to help describe her experience with MBO. If the patient agreed, consent was obtained and the semi-structured interview was performed in person and recorded by a single interviewer (CH). The timing of all interviews was 1 to 2 days prior to anticipated discharge, after most of the decision-making and counseling had occurred. If patients were readmitted during the study period, they could be reinterviewed regarding their post-discharge experience. All interviews were recorded and transcribed, with all patient identifiers and treating institutions removed prior to analysis.
For physician interviews, a list of local gynecologic oncologists was collected from a department contact list. A standardized email was sent to 27 gynecologic oncologists, and a positive response was considered consent to participate. Semi-structured interviews were then scheduled and conducted via phone or in person by a single interviewer (CH). After obtaining oral consent, interviews were digitally recorded and subsequently transcribed, either by the interviewer or professionally. All identifiers were subsequently removed from transcripts.
The interview guides were created after a literature review, and follow-up questions were revised iteratively in response to the focus of previous interviews (Suppl 1). The patient interview guide was designed to assess patient symptoms and needs during admission for MBO, decisions made during the admission as well as since diagnosis, communication with her doctor, and plans after discharge. The physician guide was designed to explore physicians' approaches to the care of patients admitted with MBO from recurrent or progressive gynecologic cancer. Questions addressed standard management practices, patient concerns, comfort with end of life discussions, and personal toll of managing MBO. The backbone of the interview guides did not change over the course of the interviews but follow-up questions and clarifications were adjusted based on each individual interview to allow interviewees to highlight areas that they prioritized.

Data analysis
Interviews were digitally recorded and transcribed. QDAMiner qualitative analysis software (Provalis Research, Montreal, Canada) was used to organize, store, and compare notes. The interviews were then systematically analyzed for themes using the framework analysis, which uses a constant comparative technique across interviews to refine themes as new data is analyzed throughout the process [9][10][11]. Framework analysis uses both deductive codes and themes, predetermined by prior research, and open coding, which is developed and refined from the data itself, and these themes are then organized and connections are created based on the data abstracted from interviews to develop an overarching theory [9][10][11][12].
Predetermined objectives included (1) to compare the priorities of physicians and patients regarding symptom control, home support, and end of life concerns; (2) to highlight aspects of patient-doctor communication during the time of MBO that functioned well and should be reinforced; and (3) to elicit communication strategies from physicians for difficult conversations, such as diagnosis of MBO. The first 3-4 interviews were coded immediately to help inform subsequent interviews on areas requiring clarification, and analysis continued throughout the interviews to establish when theme saturation was reached. Data dictionaries were developed for the patient and physician interviews and iteratively revised by 2 co-investigators (CH and NL for physician interviews, CH and FH for patient interviews). Codes were reviewed for concordance and differences in coding were resolved by mutual agreement. While there was overlap in the data dictionaries, creating different dictionaries for patients and physicians provided more flexibility to highlight unique aspects of each interviewed group.
The total number of interviews required was determined by thematic saturation. Once interviews stopped providing new themes or adding to the theories developed, no further interviews were conducted. Thematic saturation was reached after 11-13 interviews for both patients and physicians, and 2-3 further interviews were conducted for each group to ensure that no further themes emerged.

Results
Twenty patients were approached for interviews, and 14 agreed to proceed ( Table 1). The most common reason for refusal was not feeling well enough for the interview. Five of the six patients who declined to be interviewed identified as Black. More than half of patients who agreed to participate identified as White (53%), and most had ovarian/ peritoneal cancer. Age ranged from 36 to 81 years of age. Eleven patients were married. Thirteen participants were interviewed at the time of a first admission for MBO (PT 2 had a remote history of an MBO). PT 5 was readmitted and had surgery days after her first discharge; information from a second interview was combined with the first interview. Interview length ranged from 20 to 90 min.
For the physician interviews, 27 physicians were contacted via a standardized email. Absence of response was followed up with an attempt at contacting by phone. Six were never reachable, one responded and declined participation, five were willing but unable to schedule the interview, and 15 gynecologic oncologists were interviewed ( Table 2). All 15 gynecologic oncologists provided both surgery and chemotherapy for their patients.

Discrepancies in approaches and priorities
Several differences emerged in the patient experience and physician approaches to MBO (see Table 3 for select quotations). These can be categorized as differences in patient's Table 1 Characteristics of participating patients admitted for MBO from recurrent/progressive gynecologic cancer (n = 14) 1 Collected from the medical record 2 Self-identified 3 Patients treated with medications for symptom control, with or without nasogastric tube placement, without further intervention

Immediate goals
In discussing their goals for the MBO management, patients repeatedly mentioned a next, future line of chemotherapy. Physicians, however, emphasized the fact that an MBO was an end-of-life diagnosis with limited, if any, chemotherapy options. Additionally, when physicians were asked what they believed were patients' primary concerns at the time of MBO, almost all reported nutrition as foremost. Participating patients did mention nutrition but tended to emphasize the benefits of being at home.

Addressing end-of-life conversations
Patients had a wide range of responses to questions regarding end-of-life conversations. In particular, young patients with small children (PT 6, PT 9) had the shortest interviews and focused on next treatments and survival. They were not ready to discuss end-of-life planning. On the other hand, some patients who were diagnosed 5-10 years before their interview (such as PT 2, PT 10, and PT 14) seemed to accept the prognosis associated with MBO and had already delineated specific end-of-life priorities. Almost all participating physicians described the importance of end-of-life and goals-of-care conversations with patients admitted for MBO.

Communication-amount and style
Participants also brought up several differences in communication. The amount of information was a frequent theme; many patients voiced frustration with the paucity of information they received about this new diagnosis, while physicians struggled with how to prognosticate and provide counseling in the setting of so much uncertainty. In terms of style, patients valued positive thinking, while many physicians were worried about keeping expectations realistic.

Decision-making
Lastly, the experience of decision-making was very different for patients and physicians. Patient participants were asked specifically what decisions they had made about their care during their hospitalization; most patients denied having made any. Many commented that they deferred to their physicians' recommendations, and some even reported hoping they would not have to make decisions. Participating physicians wrestled with the pros and cons of medical managements as well as whether to take a patient to surgery. They noted that the amount of uncertainty about risks and benefits of all options made counseling and decision-making difficult at the time of MBO.

Factors that facilitate patient-physician communication
Patients and physicians did agree on many factors that positively influenced patient-doctor communication at the time of MBO diagnosis (see Table 4 for select quotations). First, a patient's trust of her physician was a common factor that allowed patients to understand and accept the diagnosis. Second, both patients and physicians reported that patients "being and feeling heard" improved a physician's ability to counsel the patient. The responsibility was not only on the physicians; some patients blamed themselves for not expressing themselves clearly or loudly enough, and when asked what they recommended to other women in their situation, many patient participants suggested speaking up more for themselves. Third, patients and physicians valued corroboration of information and validation of decision-making. For patients, this took the form of second opinions, personal research, and discussions with family. Physicians consulted other services (most commonly general surgery or palliative care) and stressed the importance of discussing MBO patients with colleagues. Lastly, both patients and physicians noted that time was important, both time for counseling during the acute MBO episode, and the length of time to develop a relationship with the patient prior to their admission for MBO. Time with the patient did prove to be a double-edged sword for some physicians: although a long-term relationship with the patient could make conversations easier, it came with more personal sadness and distress during goals of care discussion. Some physicians recognized that this  I think that actually the majority of patients when they come in for their first malignant bowel obstruction really don't know what it means. I'm actually kind of surprised that they really don't understand how bad of a diagnosis it can be. Across the board, no matter how educated or how much experience they have the health care system, they think, I can beat this, this is overcome-able. (MD2) I can think of one patient who was unique in that she was really direct and was not afraid to ask the next question. I think most patients, even if they get it, they're afraid to ask. (MD1) Communication style Positive thinking Keeping expectations realistic closeness might also introduce bias into the decision-making process.

Communication strategies
The participating physicians noted multiple communication strategies for difficult conversations ( Table 5). The strategies depended on physician preferences, but many also reported adapting their approach to individual clinical situations. Physicians overall were divided between offering options and always having a more directed recommendation. When multiple options appeared equivalent, they were more likely to offer options. At the same time, many physicians felt strongly that it was their professional duty as a physician to provide a recommendation. Overall, they did not find that providing statistical data was helpful to assist patient decisions, but some used best-and worse-case scenarios to put potential outcomes in perspective.

Discussion
An MBO for patients with gynecologic cancers is not uncommon and is devastating for many patients. It is also challenging personally and professionally for the physicians caring for them. It results in high symptom burden and need for increased supportive care, and forces a reevaluation of the treatment plan, an acknowledgement of an uncertain prognosis, and an exploration of end-of-life preferences. This qualitative study highlighted some differences in the priorities and concerns brought up by patients and gynecologic oncologists regarding MBO. We hope that acknowledging these differences can help physicians lead more patient-centered discussions. The widest gap between the patients and physicians' approaches to MBO was in expectations for future treatments. Many patients were very eager to get back on chemotherapy, and at the time of discharge, almost all interviewed patients reported plans to have more chemotherapy. Yet interviewed physicians opined that chemotherapy after MBO was unlikely to be beneficial. The interviewed physicians were speaking generally, not about the interviewed patients; however, there was an overall consensus about the lack of utility of chemotherapy for patients with MBO. Potentially, the patients convinced reticent physicians to give chemotherapy a try. Alternatively, their physicians did believe that there could be some benefit from chemotherapy for these specific patients based on other factors, such as prior treatment regimens or disease-free intervals. Our patient population was admittedly early in their course, in that almost all participating patients were diagnosed with their first MBO episode and had good performance status and symptom control. Only one participant was discharged with hospice. In fact, retrospective studies have found that  and diagnosis, as well as frequent check-ins after being discharged home [14]. A qualitative analysis showed that patients valued these interventions, and their daily concerns and changes in symptoms could be managed outpatient by an expert team, to prioritize their ability to be home [6].
Patients' worry about acquiring information manifested in two ways: dissatisfaction with the amount of physician communication and, more commonly, frustration at diagnostic or prognostic uncertainty of their MBO. Uncertainty is a common concern for cancer patients, and it has previously been associated with lower quality of life throughout the course of cancer treatment [15]. Participating physicians also struggled with managing uncertainty in counseling and discussion, and uncertainty is a known barrier to difficult conversations [16]. Studies suggest providing patient education can counteract some of the negative effects of prognostic uncertainty [17], but participating physicians did not find that providing statistical data was helpful for patients. Determining how to provide Table 5 Communication strategies used by physicians in the setting of difficult conversations associated with MBO diagnosis Shared decision-making A lot of it is about whether they have realistic understanding about what they may end up with, that they could end up with short bowel, and could end up with an ostomy or with chronic diarrhea and be IV dependent. … And when people have said to me, I totally get it, but I'm willing to take that risk, and I change my mind. (MD9) I usually try to put myself in their shoes and figure out what I would do. But the problem with that paradigm is that my patients are not me. But I usually do tell them what I think is best, and I tell them what we're considering, and if they feel strongly about an option that maybe I wouldn't do for myself but I think is reasonable, then I would do it. (MD12) When somebody's looking you in the eyes and saying, "How can you not offer me this thing, why won't you give it to me?" … I don't think it's right, but I don't think it's outside the standard of care to offer it to you, and if you want it and you understand the risks, I don't feel like it's my place to decide. (MD5) Providing options I give options. I let them think over the options, but then I tell them, "I'm going to come back and tell you what my preferred option is and why." … Sometimes when you come back and you ask them, they pick the option that you'd prefer, so it saves you some time. When they don't pick the option you prefer, that takes a little more time. And I have had family members come back and say: last time we were in you said that this would not be a good sign, that this is a big step back. (MD2) patient-centered information about MBO may help bridge this communication gap between patients and physicians. Some physicians used best case-worst case scenarios in their discussions. Surgeon and ethicist Margaret Schwarze suggests this as a strategy to communicate with patients in the face of uncertainty. She describes using best and worst case scenarios related to a patient's diagnosis or upcoming procedure, then discussing the most likely scenario based on surgeon experience and the clinical facts of the case [18]. Rather than depending on statistics, the scenarios provide anchors for likely and possible outcomes, giving a range and managing expectations about whether returning to a prior baseline is an option. Scenarios allow a description of what is known, and what is uncertain. It allows personalization to a patient's specific case and preferences. In light of the physicians' concerns over use of numerical data, best-worst case scenarios can provide patient education without being overwhelming.
As was previously published, most participating gynecologic oncologists feel comfortable having these goal-of-care discussions, although they are frequently conducted in conjunction with palliative care specialists [19]. Almost all participating physicians reported planning end-of-life conversations for patients with MBO, but not all patients were ready for such difficult discussions. Training during and after fellowship can help support physicians through the challenging process. Tips from well validated and accessible communication skills programs such as VitalTalk can provide strategies to address difficult aspects of the conversation (https:// www. vital talk. org/ clini cians/). Our study highlights the particular difficulty in addressing these concerns for recently diagnosed, young patients, and suggests early involvement with palliative care specialists for these patients in particular may help provide support for physicians and patients in symptom management, family support, and eventual planning for end-of-life care.
The findings described in this paper have led us to believe that a structured communication tool could address the struggles physicians and patients describe, including managing uncertainty and communication of information. Decision aids have been shown to help patients have a more active role in decision-making and feel better informed about their diagnosis and their values [20]. Decision aids were found to successfully provide patient education in the setting of suboptimally debulked ovarian cancer [21], but a review of the literature revealed no decision aids created for MBO. While a "decision aid" may seem out of place in a setting such as MBO where many patients did not remember making decisions and most physicians did not feel they had many options to give, it can help highlight decisions that are being assumed and bridge communication gaps highlighted in our interviews.
A decision/communication aid in the setting of MBO should prompt physicians and patients to better define realistic patient goals. In particular, many of our participants wanted to go home; eliciting home needs and providing follow-up could help empower patients after discharge. An understanding of expectations with chemotherapy after MBO could also be clarified. A plan for worsening of symptoms should be discussed. It should address the possibility and risks versus benefits of future treatments, and events that would require stopping chemotherapy. We plan to use these findings to develop and pilot a decision/communication aid for patients with MBO.
Our study is not without limitations. Both physicians and patients were limited to a single urban Midwest geographic area. Additionally, most patients who agreed to the interview were admitted for their first admission for MBO and had good symptom control; they may not be as representative of patients with more advanced or recurrent MBO. Only women with gynecologic cancer were included in this study, and they may not be representative of men and women with non-gynecologic cancers, or at other time courses in their disease. It is unknown whether patients or physicians who declined to participate in the interviews would have provided different insight. Many of the physicians interviewed, as well as three of the patients, were known to the interviewer outside of the study. Physicians were questioned in general terms about patients with MBO, while participating patients were focused on their individual experience, which can limit the comparisons between the two interview groups. As a qualitative study, it is designed to bring further questions that can be explored. Other perspectives that would be interesting to investigate include those of bedside nursing, family members, home health, palliative care specialists, and hospice teams.
Our study's strengths include the ability to interview both patients and gynecologic oncologists to address different perspectives of the challenging topic. The interviews were conducted by one interviewer for consistency, and she was not clinically involved in the majority of the care of the patients interviewed. The framework method allowed us to start with themes from the literature while simultaneously developing codes and themes coming from the interviews themselves.

Conclusion
Interviews with women with MBO and gynecologic oncologists who care for women with MBO revealed discrepancies in approach between the two groups, but also factors that facilitate the patient-physician communication that can be further developed. A patient-centered communication guide for physicians and patients may be helpful to trigger discussions about aspects of care that patients highlighted, such as nutrition and timing to discharge home. It could also leave patients feeling better informed in the face of uncertainty and physicians better supported to provide information, which were themes in both groups of interviewees. Further research to develop a formal communication aid for physicians and women with MBO in the setting of recurrent gynecologic cancer and to tailor it to specific patient and physician populations and hospital resources is needed.
Author contribution CH, FH, JC, SDY, and NL all contributed to the project development and study design. CH conducted the interviews and some of the transcription. CH, FH, and NL shaped the data dictionary and analyzed the transcripts. CH, FH, JC, SDY, and NL contributed to writing the manuscript.
Funding Partial financial support was received from Bears Care, the charitable arm of the Chicago Bears Football Club (CH, SDY).

Data availability Not applicable.
Code availability Not applicable.

Declarations
Ethics approval The patient interviews were approved by the IRBs of the University of Chicago and Northshore University Health System. Physician interviews were granted IRB exemption by the University of Chicago.

Consent to participate
Patients signed a standard IRB-approved consent prior to participation, which included agreeing to audio recording. Physician consent was confirmed by a positive response to a standard email, and verbal consent was elicited for audio recordings.