Cancer Patients’ Preferences for Psychological Counselling: A Cross-Sectional Study Using Full-Prole Conjoint Analysis in Japan

Background: Psychological counselling is an effective non-pharmacological intervention for cancer patients facing psychological distress. However, there are barriers to applying counselling in clinical settings because some cancer patients feel that it is a burden and subsequently drop out. This study assessed cancer patients’ preferences for psychological counselling using a conjoint analysis approach. Methods: We conducted a single-center, cross-sectional study using a self-reported questionnaire. The data were collected from outpatients and inpatients at the departments of Respiratory Medicine, Gastrointestinal Surgery, Female Pelvic Surgery, and Head and Neck Surgery from a university hospital between March 2018 and March 2020. The questionnaire, developed from an interview with a clinical psychologist, asked for patients’ preferences for counselling with 48 scenarios according to a combination of type, length, price, and frequency. We conducted a conjoint analysis, calculating the relative importance and part-worth value of each factor. Results: According to the conjoint analysis modeling, cancer patients consider the type as the most important factor when deciding whether to undergo counselling (relative importance: 37.5%), followed by frequency (23.4%), price (19.6%), and length (19.5%). Cancer patients value life reviews highly among the four counselling types. However, women and younger patients valued counselling for emotional control the highest, with advanced cancer patients valuing counselling for problem solving the most. Conclusions: The results reveal that cancer patients consider the type and content of counselling to be the most important factors in the selection process, although preferences depend on individual characteristics. This is the rst study to clarify cancer patients’ preferences for psychological counselling using conjoint analysis and proposes a new economical method in the psycho-oncology eld.

Psychological counselling is the most useful non-pharmacological treatment to help cancer patients in managing their psychological distress [8][9][10]. Various types of counselling approaches have been developed for use among cancer patients. Previous studies have outlined the e cacy of numerous counselling techniques for cancer patients, such as supportive counselling, cognitive behavioral therapy, and psychoanalytic therapy [8][9][10]. Additionally, cognitive behavioral therapy includes problem-solving therapy, behavioral interventions, and mindfulness cognitive interventions [11][12][13].
Psychological counselling has a clear effect on psychological distress in cancer patients. However, there are barriers to applying counselling in a clinical setting. First, meta-analyses demonstrate that there is heterogeneity between studies due to wide variations among the psychotherapeutic interventions studies [10].
Second, the e cacy of psychological counselling depends on the clinical communication and utilization of hybrid skills of the practicing psychologists [14][15][16]. Third, few studies have described any adverse event data of psychological counselling, although the dropout rate is known to be high according to previous research [17,18]. Therefore, an understanding of cancer patients' needs around counselling is still needed to provide a more effective treatment method for their psychological distress.
Therefore, this study aimed to assess cancer patients' preferences for psychological counselling using a conjoint analysis approach. We studied the adequacy of the application of an economic approach to clarify cancer patients' counselling preferences and rst conducted a single-centre trials in Japan. This study seeks to provide information to allow for an improved consultation rate of psychological counselling among cancer patients.

Study setting and participants
We conducted a single-center, cross-sectional study using self-reported questionnaires. Patients were recruited from outpatient and inpatient units from a university hospital. The data were collected in the departments of Respiratory Medicine, Gastrointestinal Surgery, Femoral Surgery, and Head and Neck Surgery using selfreported questionnaires between March 2018 and March 2020. We included patients who (a) had cancer diagnoses and (b) were between the ages of 20 and 84 years old. The following patients were excluded from this study: (a) those who declined to participate, (b) those who were diagnosed by their attending doctor with severe physical, mental, or cognitive problems (e.g., serious adverse events, delirium, or dementia), and (c) those who had insu cient literacy skills.

Identifying Attributes
We conducted a preference elicitation study using semi-structured interviews with ten clinical psychologists who have experience in providing psychological counselling for cancer patients. This qualitative approach has been used in other preference studies to identify key factors [19,20]. The interview was composed of the same counselling features that the clinical psychologists would usually provide with their clients, including the following factors: (1) type, (2) content, (3) length, (4) price, and (5) frequency. Each interview's audio was recorded, transcribed, and analyzed using content analysis. The result of this qualitative study suggested that length was shorter and the price was lower in cancer palliative settings than those in a psychiatric setting.
Based on the psychologist interviews, Table 1 presents the nal attributes and levels chosen for the conjoint analysis, with each description used in the questionnaire survey instrument. From the interviews, we combined the type and content of counselling sessions into a single theme, identifying the nal four attributes and levels used in our analysis (Table 1). For the interviews, we used 4,000 Japanese yen as a benchmark for the price of counselling, while setting 1,500 Japanese yen as a comparative health treatment price referring to Japanese insurance institutions in cancer palliative settings. In addition, we used 30 minutes as the benchmark in time spent within private room counselling, while setting 10 minutes as the comparative time spent for bedside counselling sessions in a cancer palliative setting. Additionally, a researcher (K.I.) translated and created the English versions of the attribute and level descriptions of the counselling features utilized in this study. A master's student translator, who has been living in English-speaking countries for the past four years, then back-translated the English version into Japanese.

Clinical characteristics
We collected both demographic and clinical information from the self-completed questionnaires and the medical records. First, we included the following data: sex, age, place or recruitment, types of cancer, cancer
We de ned the age between 20 and 59 years old as younger, and stage III or IV cancer as advanced cancer.
Second, we asked about patients' desire to attempt psychological counselling using the 5-point Likert scale questionnaire. We originally developed this question for this study.

Sample Size
The sample size calculation for a conjoint analysis is generally empirically determined because it is dependent on several conditions, such as the number of attributes and levels, as well as the complications of the given task. The sample size of the conjoint analysis ranges from 100 to 300 participants in previous healthcare studies [21]. Therefore, the sample size needed for this study was estimated to be almost 200 participants.

Statistical analyses
First, we summarized the characteristics of both the participants and hospitals utilizing standard descriptive statistics. Second, we conducted the conjoint analysis and calculated the relative importance and part-worth value of each of the counselling factors. Third, we conducted a subgroup conjoint analysis among participants who were women (n = 86), of a younger age (n = 49), and had advanced cancer (n = 91). All data were analyzed using R version 4.0.2 with the R packages "dplyr," "tidyr," "readr," "readxl," and "conjoint."

Preferences for clinical counselling
According to the conjoint analysis model (Fig. 2), the participating patients considered the type of counselling as the most important preference factor (relative importance: 37.5%), followed by frequency (23.4%), price (19.6%), and length (19.5%). The subgroup analysis also indicated that counselling type is the most important among women, patients of a younger age, and those at an advanced cancer stage (35.7%, 30.7%, and 38.2% respectively; Table 3). The participating cancer patients valued life reviews highly among the four types of counselling when compared to problem-solving, emotional control, or emotional expression (Fig. 3). However, women and younger age patients valued counselling for emotional control the most, with advanced cancer patients valuing counselling for problem solving the highest (Table 3). In addition, all of the participating cancer patients valued a 30-minute duration over that of 10 minutes, as well as a price of 1,500 Japanese yen instead of 4,000 Japanese yen (Table 3).

Discussion
Our study yielded two major ndings. First, we observed that the type of counselling was more critical for its selection among cancer patients than was price, length, or frequency. Second, our results demonstrated, in terms of counselling type, that cancer patients prefer life review therapy compared to those involving problem solving, emotional control, or emotional expression. This study was the rst notable attempt to explore cancer patients' needs and preferences for counselling in terms of medical services.
One important consequence from this study is that cancer patients choose whether to participate in psychological counselling while emphasising the intervention's content. The issues that cancer patients seek support for are clear among this population, although their perceived needs or worries are often quite different from one another [22,23]. In addition, adjustment disorder is the most common psychiatric diagnosis in cancer patients [24,25]. Cancer patients often experience clear and speci c stressors rather than any original mood symptoms. However, this result may have been affected by either the chosen attributes or levels of the included counselling features. The maximum possible counselling price, 4,000 Japanese yen, is affordable for cancer patients in Japan as it is a nancially secure country. Furthermore, the minimum counselling length, 10 minutes, was found to be enough time for the participants of this study who did not have a strong desire to attempt counselling. Additionally, the characteristics of the counselor or counselling room may be a more critical component for cancer patients than the type, length, price, or frequency of counselling. Therefore, this study suggests that the type of counselling is important for cancer patients when choosing whether to undergo this treatment method if no other conditions are being considered.
Another nding was that life review is the most preferred counselling type among cancer patients. Life review provides patients an opportunity to construct positive thoughts about their lives, with it being one of the most effective non-pharmacological interventions, in addition to cognitive-behavioral therapy or mindfulness-based stress reduction [26,27]. Most patients will face various inevitable problems related to their disease and will thus require an alternative treatment strategy to maintain their mental health even if they are unfamiliar with psychological interventions. Most of the participants in our study were also at an elderly or retiring age, meaning that they were likely contemplating the end of their lives in some form. Therefore, life review undoubtedly contributes to the reduction of the psychological distress experienced by elderly cancer patients [28]. In this regard, however, the results of our subgroup analyses revealed that cancer patients' preferences for counselling depending on their characteristics, such as their gender, age, or the current state of their disease. Women and younger patients tend to prefer counselling covering emotional control, according to this study's ndings. This indicates that these patients hope to gain a sense of control over their emotions, especially when one considers the fact that women and younger patients face a similar risk for psychological distress, including problems like anxiety or depression when dealing with cancer [29,30]. It has been found that most young adult cancer patients and survivors need psychosocial support in the form of health information or in terms of their sexuality problems [31], so that they may focus on resolving their present feelings rather than reviewing those from their past. Conversely, our results also reveal that advanced cancer patients prefer problem-solving in terms of the counselling type. Another study demonstrated patients with poor performance statuses have unmet needs in most domains involving physical, psychological, or social problems [32]. Therefore, advanced cancer patients may hope to spend their efforts resolving the physical or social problems at hand, rather than trying to control their psychological distress or review their lives.
In addition, there were no differences found in patients' needs for counselling in all subgroups. The participating cancer patients naturally chose the less expensive counselling option; however, it is di cult to provide counselling at 1,500 Japanese yen given therapists' labor costs. The Japanese government created a national license, entitled "Certi ed Public Psychologist," for all clinical psychologists in 2019 [33]. The results of our study suggest that there is a compelling need for applied counselling as a healthcare treatment plan for cancer patients.

Study limitations
Our study has two important limitations. First, there were limitations to the conjoint analysis. As previously noted in the discussion, the type of counselling could not have been the most critical factor in patients' selection of counselling across all participants because this varies depending on the options available. However, we selected the attributes and levels of counselling features in our conjoint analysis based on interview research gained by Japanese clinical psychologists in cancer palliative settings. Second, selection bias might have occurred because of our use of a single-center design in Japan. In this study, the number of breast cancer cases was smaller than that in previous studies. Thus, this may impact the generalizability of our ndings. In addition, the number of patients who had an experience with or desire to undergo psychological counselling was especially small in this hospital. However, this was the rst study to focus on cancer patients' needs for counselling.

Conclusions
Our study indicates the possibility that conjoint analysis can apply not to just economic studies but also cancer patients' preference studies. The nding is that most cancer patients consider the type or content of counselling as the most important factor when determining whether or not they wish to receive this treatment form, although their preferences for counselling do depend on their characteristics. This was the rst study to clarify cancer patients' preferences for psychological counselling, despite its few methodological limitations.

Abbreviations
No Abbreviations.

Declarations
Ethics approval and consent to participate All the procedures were in accordance with the ethical standards of the responsible committees on human experimentation (institutional and national) and with the Helsinki Declaration. The study was approved by the independent institutional review board (IRB) of the Tokyo Medical and Dental University (M2017-226).
Informed consent was obtained, in writing, from all participants after they had received verbal explanations of the study procedure. Participants answered anonymous questionnaires and were identi ed by code numbers.

Consent for publication
Not applicable.
Availability of data and materials The datasets during and/or analyzed during the current study are available from the corresponding author upon reasonable request.  Figure 1 Flow diagram of the study participants Figure 2 Relative importance of the different attributes Figure 3 Part worth values of the counselling factors