Analysis of Patient-Physician Concordance in the Understanding of Chemotherapy Treatment Plans Among Patients With Cancer

Key Points Question Do patients with cancer understand the treatment plans to which they consent? Findings In this cross-sectional study of 151 adult patients with cancer and 20 treating physicians only 13.7% of patients had full concordance with their physician regarding aspects of chemotherapy treatment plans. College or advanced degrees and a family history of cancer were associated with better understanding of treatment plans. Meaning These findings suggest that more effort and time should be invested in enhancing the understanding of chemotherapy treatment plans among patients with lower educational levels and/or no family history of cancer.


Introduction
The journey of a patient's treatment starts with effective patient-physician communication. Effective communication is associated with high-quality health care. 1 Moreover, understanding a patient's concerns and preferences with regard to treatment options can aid physicians in identifying misconceptions and lead to better medical decisions. 2 Nevertheless, effective communication between physicians and patients regarding treatment plans is not easily achieved. 3 This challenge is more pronounced in patients with cancer because of the complexity of care needed. 4 The treatment plans of patients with cancer involve multiple modalities, for which oncologists need to explain the therapy goals, duration, and expected complications of treatment. Therefore, effective communication between patients with cancer and their treating physicians is important to ensure patient adherence to treatment and achieve better outcomes.
It is also possible to change treatment plans for such patients according to certain factors, such as tumor response to treatment and toxic effects that may arise from chemotherapy. 5,6 Discordance in the understanding of the treatment plan between patients and their physicians is not uncommon in the Western world. In 1 study that evaluated hospitalized patients' understanding of their treatment plans, patients perceived themselves to be fully aware, yet they were in disagreement with their physicians on some aspects of the treatment plans. The study assessed the knowledge of patients at hospital discharge, and no more than one-half of them had accurate information about each aspect of their conditions, which included their diagnoses, medications, and adverse effects of treatment. 7 Patient education in clinical practice has been associated with substantial positive outcomes.
Interventional studies that included targeted health education following a comprehensive assessment of patients' understanding have indicated positive results. 8,9 It has been reported that therapeutic education is associated with a decrease in the number of hospitalizations for bronchial asthma and diabetic coma as well as a reduction in the number of lower limb amputations. 10 Other studies have indicated that physicians and trainees frequently overestimate patients' understanding of their treatment plans at hospital discharge and do not recognize patients' low health literacy. 11,12 Although only 57.0% of hospitalized patients reported that they understood the potential adverse effects of their medications on discharge, physicians believed that 89% of their patients understood these effects.
To our knowledge, no study has been carried out in Saudi Arabia regarding this topic. The provision of health care in Saudi Arabia has some unique features that may add to this challenge.
These features include the cultural understanding of cancer and the stigma surrounding certain types of cancer, the lack of common medical terminology that patients can understand (which may also result from a patient's lack of health education), and the fact that a substantial number of health care professionals do not speak Arabic. 13 Furthermore, the strong ties between Saudi family members can add another challenge regarding the expected level of care. The families of patients with terminal cancer usually expect curative treatments even after any realistic hope of a cure is gone. 14 The goal of this study was to investigate the concordance between patients and their treating oncologists in the understanding of chemotherapy treatment plans and to investigate the potential patient-related and physician-related factors in this concordance.

Study Design and Setting
An interview-based cross-sectional study was conducted at the Princess Noorah Oncology Center in King Abdulaziz Medical City (Jeddah, Saudi Arabia), which has a total capacity of 751 beds and includes medical and surgical departments in addition to the oncology center. This tertiary medical center is one of the major oncology centers in Saudi Arabia and provides care to the Saudi population and other eligible patients in the western region of Saudi Arabia. The oncology center has 108 beds and is mainly composed of 6 departments that include adult medical oncology, gynecological oncology, adult and pediatric hematology and bone marrow transplant, pediatric oncology, radiation oncology, and palliative care. This study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies, and it was approved by the institutional review board of the King Abdullah International Medical Research Center.
Participants were asked to voluntarily participate in the study. An approved informed consent form designed by the research center was given to all participants before the interview. Participants' identities were confidential, and each participant was assigned a generic serial number that was not linked to the medical record number. Participants had the right to withdraw from the study at any time. The collected data were kept confidential and protected from any access by a third party. The data were stored and the passwords protected in the workplace computer, to which only the authors had access.
The study used a consecutive sampling method for selecting the participants between October 4, 2017, and November 8, 2018. The data collection method was an interview-based structured questionnaire, for which patients and physicians were interviewed separately. Patients' responses were compared with physicians' responses to assess the level of concordance. Data were analyzed from November 15 to December 20, 2018.

Participants and Sample Size
The study included Saudi adult patients older than 18 years who were scheduled to receive a therapeutic pharmaceutical cancer treatment that required them to sign an informed consent document. Patients who had a personal history of cancer or were unwilling to be involved in the decision-making process were excluded (Figure).
The annual number of patients who receive first-line therapeutic pharmaceutical interventions at the Princess Noorah Oncology Center is approximately 877. 15 The study used a 95% CI with a 5% margin of error and an assumed response distribution of 50%. The minimum required sample was calculated to be 268 patients.

Questionnaire
The study used a unified questionnaire as a data collection tool during patient and physician interviews (eMethods in the Supplement). The unified questionnaire ensured that answers were consistently reported during interviews performed by multiple data collectors (H.A., A. Absi, A. Alghamdi, and M.A.). Patients' and physicians' interviews were conducted separately.
The questionnaire was composed of 5 parts. The first part addressed the demographic characteristics of each patient and treating oncologist. Patients were asked about their age, sex, educational level, type of cancer, and family history of cancer. Physicians were asked about their age, Absi and an independent oncologist). To eliminate the language barrier among patients, the questionnaire was translated into Arabic. A back-translation into English was also performed by a language expert. To ensure the reliability of the questionnaire, a pilot study was conducted on 25 patients. Any confusion or misunderstanding regarding the questions was identified, and the questions were modified appropriately. The final revised version was then used in the study.

Variables and Bias
The study explored the association between patients' understanding of their treatment plans and their treating physicians' sex, age group, primary language, nationality, and practice location (within or outside of Saudi Arabia). The demographic data of patients and physicians were obtained from the ARIA database, and complementary data were obtained through an interview questionnaire to ensure consistency in the responses of patients and physicians. All variables were qualitative and were coded for storage and analysis using IBM SPSS Statistics, version 23 (IBM SPSS). Qualitative variables were reported as frequencies and percentages. Patientphysician concordance level in therapy goals, duration, outcome, and toxic effects were evaluated based on the patients' sex, age group, educational level, family history, and type of cancer.
To minimize recall bias, the study included patients who were scheduled to receive a therapeutic pharmaceutical cancer treatment that required them to sign an informed consent document after speaking with their treating physicians. Patient and physician interviews were then conducted separately within 24 hours after the patient signed the chemotherapy consent form.
Another potential bias was the fact that data collectors reported the patients' answers. Therefore, the patient interview was conducted first to blind the data collector to the expected answer that was discussed by the physician.

Statistical Methods
Analyses of descriptive statistics were performed and reported as proportions and percentages for categorical variables and means (SDs) for continuous variables. Patients' understanding of treatment plans was evaluated by comparing their answers with those of their treating physicians, and understanding was indicated by the level of concordance between the 2 sets of answers. The overall level of concordance was assessed by the number of correctly answered questions. If all 4 treatment plan-related questions were answered correctly, the patient-physician concordance level was labeled full. If 1 or more questions were answered incorrectly, the concordance was considered partial. Incorrect answers on all questions indicated full discordance.
The association of sociodemographic factors (among patients and physicians) and family history of cancer (among patients) with overall patient-physician concordance in the understanding of the treatment plan was assessed using a χ 2 test with an analysis of the adjusted residual (AR) and a Fisher exact test. Tests were 2-sided and unpaired, and the statistical significance threshold was P < .05.  Table 1). Twenty treating oncologists were interviewed, of whom 14 (70.0%) were men and 6 (30.0%) were women. Arabic was the primary language spoken by 19 physicians (95.0%), and 19 physicians (95.0%) had practiced medicine at some point outside of Saudi Arabia ( Table 2). A total of 87 patients (57.6%) were receiving care from medical oncology teams, and 59 patients (39.1%) were receiving care from hematology teams. Seventy-seven patients (51.0%) were interviewed in an hospitalization. All 20 of the treating physicians participated and were interviewed more than once because they provided care for multiple patients in the study.

Results
No patients had complete discordance with their treating physicians on all aspects of their treatment plans. Therefore, patients were divided into 2 groups based on the level of concordance with their physicians. The first group consisted of those with concordance on 1 or more, but not all, aspects. This group represented most of our sample (131 patients [86.2%]). Those who had full concordance with their physicians on all aspects were included in the second group and accounted for 20 patients (13.7%). Table 3, the highest patient-physician concordance rate was associated with the type of cancer, for which 123 patients (81.5%) correctly identified their primary cancer diagnosis.

As illustrated in
In contrast, a notable patient-physician discordance was observed in the duration of the Adjusted residual values were used to identify statistical significance. A statistically significant difference was found between overall patient-physician concordance across different age groups of physicians (χ 2 1 = 5.84; P = .02). Older physicians (aged >40 years) were more likely to achieve higher rates of full concordance with their patients compared with younger physicians ( Table 4). In contrast, patients older than 60 years were more likely to have partial rather than full concordance with their physicians compared with patients younger than 60 years (37 patients vs 0 patients, respectively; χ 2 1 = 5.84; P = .008), with an AR of 2.7. In addition, the level of overall patient-physician concordance differed significantly across the 3 levels of patient education (χ 2 1 = 17.73; P < .001).  Patients with college or advanced degrees were more likely to have full concordance with their physicians (AR = 4.1), while patients with less than a high school education were more likely to have only partial concordance (AR = 3.0). A family history of cancer was associated with a greater likelihood of full patient-physician concordance (χ 2 1 = 15.88; P < .001).

Discussion
This study assessed the level of patients' understanding of their treatment plans based on the level of patient-physician agreement in a single oncology center. We observed variable levels of discordance depending on the aspect of the treatment plan. There are multiple explanations for this discordance that pertain to either the physicians' characteristics and the setting of the encounter or the patients' characteristics and backgrounds.
In our study, 81.5% of patients were able to correctly identify their diagnosis. However, only 13.7% of patients were in full concordance with their respective physicians in all aspects of their chemotherapy plans. In addition, 55.8% of patients were expecting shorter treatment durations than were their physicians. In another study conducted at the Mayo Clinic, patients were not in agreement with their treating physicians in multiple domains, including diagnosis, planned tests and procedures, medications, and expected date of hospital discharge. 7 Furthermore, the present study demonstrated a patient-physician concordance of 68.2% in treatment goals, which was comparable with 1 study conducted on patients with cancer in which 69% of patients were in agreement with their physicians with regard to treatment goals. 16 The remaining discordance might be attributed to multiple factors, such as the fact that physicians may have overestimated their patients' understanding of the treatment plans or may not have recognized their patients' levels of health literacy. 11 Furthermore, time constraints and the load of clinical duties might be factors in the level of physicians' engagement in discussions with patients.
Multiple studies have assessed patients' understanding of the common adverse effects of treatment plans in different settings. In a study performed in Columbus, Ohio, 135 patients were contacted 2 to 6 weeks after hospital discharge to assess their level of awareness of and adherence shared decision-making model is being increasingly implemented today. 18 An integral part of the model is the proper explanation of informed consent to ensure the patients' understanding of treatment benefits, the potential for serious adverse effects, and the requirements of further diagnostic evaluation. 19 The educational level of the patient was significantly associated with a better understanding of the treatment plan, which was manifested as a higher patient-physician concordance. Patients who had college or advanced degrees were more likely to agree with their physicians on treatment plans.
Educational level is a recognized factor in a patient's ability to provide informed consent and understand the treatment plan. An increased level of understanding of cancer was observed in those who had a higher level of education. 20 22 In addition, shared decision-making is associated with improvements in overall patient adherence to and satisfaction with cancer treatment plans. 23

Limitations
This study had several limitations. First, the interviews were conducted within 24 hours after the patient-physician encounter to minimize recall bias. However, patients who were receiving multiple treatment modalities, including chemotherapy, radiotherapy, and surgical interventions, had difficulty recalling different aspects of their treatment plans. Second, this study was performed in a single center. Therefore, larger-scale studies are necessary for a more precise estimation of patientphysician concordance and assessment of the associated factors. A collaborative study with another oncology center would allow further exploration of the potential cultural differences that may help to explain aspects of the results. Third, patient-physician agreement is subject to the constantly evolving nature of cancer, which can necessitate multiple changes in treatment management plans.
Future research should be directed at evaluating the association of the patient's cancer stage and overall prognosis with the patient's perception and understanding of treatment goals and plans.

Conclusions
Most of the patients in this study showed a suboptimal understanding of aspects of their chemotherapy plans. Higher educational levels and family histories of cancer were associated with better understanding. More effort and time should be invested in enhancing the understanding of chemotherapy plans among patients with lower educational levels and/or no family history of treatment with such therapies. In addition, a patient self-report evaluation of the understanding of chemotherapy plans could be added to the informed consent process to assess patients' level of understanding and develop a stepwise patient education program that targets those with the lowest levels of understanding.