Design and participants
This was a national cross-sectional survey conducted among Japanese health care providers in oncological and cardiovascular hospitals using self-completed questionnaires. We mailed the questionnaires to the chief physicians of respiratory and cardiovascular internal medicine of 347 designated cancer hospitals and 427 ICD specialized hospitals, asking them to deliver the questionnaires directly to oncological and cardiovascular physicians and nurses in March 2018. Designated cancer hospitals, recommended by the prefectural governments, can provide high-quality cancer treatment guaranteed by the Ministry of Health, Labour and Welfare in Japan. They provide specialized cancer treatments, establish local cooperation systems for cancer treatments, and provide consultation, support, and information for cancer patients. Additionally, ICD specialized hospitals can operate implantation of implantable cardioverter-defibrillators (ICDs).
Ethics
Our study was approved without undergoing assessment by the institutional review board at the Tokyo Medical and Dental University in Japan, as it was an anonymized non-invasive self-completed questionnaire study for health care providers. However, all procedures were in accordance with the ethical standards of the responsible committees on human experimentation (institutional and national) and with the Helsinki Declaration of 1975, as revised in 2000. All participating physicians and nurses were volunteers and were not provided with any incentive to participate. The anonymized questionnaires were individually returned from each health care provider and not specified by an individual person or facility.
Demographic and clinical characteristics
We collected demographic and clinical information from the self-completed questionnaires. First, we included the following data: sex, age, and medical license of the staff of each health care provider. Second, we included the following data: area (Hokkaido/Tohoku, Kanto/Koshinetsu, Chubu/Hokuriku, Kinki, Chugoku/Shikoku, and Kyushu/Okinawa area), hospital type (national medical center, academic medical center, general hospital except academic medical center, specialized hospital), the number of hospital beds, and the presence of a palliative care unit, palliative care team, liaison psychiatry team, palliative care physicians, psychiatrists, and psychologists at hospitals.
Outcome measures
Difficulty of providing palliative care
The Palliative Care Difficulties Scale, a 15-item self-reported scale, was developed in Japan [31]. It is a 4-point Likert-type scale ranging from 0 to 3 (overall score range: 0–42). The scale contains five factors with three items each: (1) alleviating symptoms, (2) expert support, (3) multidisciplinary communication, (4) communication with patient/family, and (5) community coordination. The reliability and validity of this measure were sufficiently supported in an earlier study [31].
Difficulty of providing end-of-life psychiatric care
We developed the following question for assessing the difficulty of providing end-of-life psychiatric care: “Do you feel challenged to provide psychiatric care for patients at their end of life?”; possible answers were “yes” or “no.”
Barriers to providing end-of-life psychiatric care
To identify barriers to providing end-of-life psychological care, we asked participants who answered “yes” in the above question: “Why do you feel challenged to provide psychological care for patients at their end of life?” and asked to submit written answers in free description.
Qualitative analyses
Content analysis was used to analyze free description data. Content analysis is an objective and systematic procedure used to draw conclusions by creating categories of data from verbatim or unstructured data [32]. We conducted a quantitative content analysis according to previous studies in palliative care settings [33, 34]. Our content analysis procedure was conducted as follows: (1) all text data were divided into thematic units, which are the units of words with one logical meaning; (2) two researchers, a clinical psychologist and a cardiovascular nurse (IK and MS, respectively) extracted all statements from the free descriptions related to the study topic, such as the barriers to providing end-of-life psychiatric care; (3) a clinical psychologist (IK), a cardiovascular nurse (MS), and two psychiatrists in the palliative care team (ME and TT) carefully conceptualized similarities and differences in the content, and defined all categories; and (4) two coders, a student of psychology and a psychiatric clinical nurse, independently determined how each thematic unit that was identified corresponded with any category. The concordance rate and kappa coefficient of the determinations of the categories were used as reliability indicators. The kappa coefficient was calculated using 20% of the data and random sampling was conducted based on the data from a standard set derived from a previous study, with more than 10% or 50 units of data [35, 36].
Statistical analyses
First, we summarized the characteristics of the participants and hospitals using standard descriptive statistics. Second, the mean difference in difficulties in providing palliative care was compared between oncological and cardiovascular hospitals using a t test, and the frequency of difficulties in providing end-of-life psychiatric care was compared between oncological and cardiovascular hospitals using a χ2 test. Third, the frequency of the thematic units that were categorized in the above content analysis was compared between health care providers in oncological and cardiovascular hospitals using a χ2 test. The significance level was set at 5%. All data were analyzed using IBM SPSS Statistics for Windows, version 24 (IBM Corp., NY, USA).