The scale was developed using a draft scale consisting of items extracted quantitatively [14] and inductively from 10 home-based patients with advanced cancer who were receiving opioids. These items were extracted deductively from previous studies [4,5,15,20]. We examined the Phase 1 content validity and Phase 2 surface validity. A refined draft scale for the validity and reliability of the Opioid Self-management Scale for Patients with Advanced Cancer (OSSA) in Phase 3 was considered.
Phase 1: Content Validity
Participants and Procedures
The inclusion criteria were 1) researchers with experience in cancer pain research and 2) physicians, pharmacists, specialists, and certified oncology nurses with direct involvement in cancer pain care. The exclusion criteria were 1) researchers with no experience in cancer pain research and 2) physicians, pharmacists, specialists, and certified cancer care nurses with no direct involvement in cancer pain care. The minimum calculated sample size was five [21]; therefore, we aimed to recruit at least five professionals from each category for a total of 30 researchers, physicians, pharmacists, and nurses as subjects.
Study information and questionnaire forms were given to the participants between August and October 2018. The variables of interest included occupation, specialist’s qualifications, years of experience in cancer pain research or care, and draft scale 1. The scale consisted of 15 items on “Understanding pain and opioids,” 12 items on “Coping with pain by taking opioids,” 3 items on “Coping with side effects of opioids,” 5 items on “Having someone that can provide support,” and 5 items on “Living in your own way while coexisting with the disease” for a total of 40 items in five subscales. The evaluation was conducted using the content validity ratio (CVR), and the subjects were asked to choose one of the following fitting responses: “Very appropriate,” “Very appropriate but expression needs to be corrected,” “There is a problem with the appropriateness,” and “Inappropriate.” The CVR was calculated using the formula:
ne indicates the number of subjects who chose “Very appropriate’’ and “Very appropriate but expression needs to be corrected”. indicates the total number of subjects [22].
Furthermore, the subjects who chose “Very appropriate but expression needs to be corrected,” “There is a problem with the appropriateness,” and “Inappropriate” were asked to propose corrections, additions, and provide feedback.
Data Analyses
Using a CVR of 0.62 as the standard, items that did not meet the standard were deleted [23]. The participants’ recommendations were discussed by two nursing researchers with experience in developing scales and five graduate students in nursing research. The items were then refined to create draft scale 2.
Phase 2: Surface Validity
Participants and Procedures
The inclusion criteria were: 1) patients with advanced cancer, 2) administration of opioids for cancer pain alleviation, 3) over 20 years old, and 4) aware of the name of their disease. The exclusion criteria were: 1) patients without advanced cancer, 2) administration of opioids for other purposes apart from cancer pain alleviation, 3) younger than 20 years old, 4) aware of the disease name, and 5) having physical, mental, and cognitive disorders according to their responses. The sample size was 10 [19].
The information document of the study and questionnaire forms were given to the participants, who were outpatients in a designated cancer hospital in October 2018 and patients receiving home-based care. Subjects’ submission of the questionnaire forms was considered consent provision. The assessed variables in the questionnaire were age, sex, performance status (PS), duration of opioid use, and draft scale 2. Draft scale 2 assessed the appropriateness of the responses, proposals for improvement, and participants’ feedback: they were asked to comment freely.
Data Analyses
Two nursing researchers with experience in developing scales and five graduate students in cancer nursing research discussed the choice of expressions for the items, refined the items, and created draft scale 3.
Phase 3: The Validity and Reliability of the OSSA
Participants and Procedures
The inclusion and exclusion criteria for Phase 3 were the same as those for surface validity. The calculated sample size was 130. Based on an α coefficient of 0.90, confidence interval of 0.05, and estimated response rate of 60% [19], 210 subjects were recruited.
The information document of the study and questionnaire forms were given to the participants, who were outpatients at three designated cancer hospitals. The questionnaires were self-administered and returned on site or via mail. Submission of questionnaire forms by patients was considered consent provision. Patients were re-tested one week later, and the forms were mailed back.
Measures
Participant Characteristics
Patient characteristics included age, sex, PS, presence or absence of a caregiver, employment form, patient history, disease evolution, treatment history, types of opioids, opioid doses used, duration of opioid use, type of pain and extent of pain (was assessed using the numeric rating scale [NRS]: most intense pain, average pain, and pain that interferes with daily life activities; and the rate of pain relief over a 24-hour period).
OSSA
Draft scale 3 was used. Each score is a Likert scale score ranging from 1 point for “No” to 5 points for “Yes.” Higher scores indicate greater self-management ability.
Self-Care Agency Questionnaire(SCAQ) [24]
There are 29 items in four subscales, including 10 items on “Ability to perform self-care operations,” 7 items on “Ability to adjust one’s own physical condition based on personal weaknesses,” 7 items on “Ability to concentrate one’s attention on self-care,” and 5 items on “Ability to receive valid support.” The scores range from 1 point for “No” to 5 points for “Yes.” A higher score indicates greater self-care ability.
Medication Adherence Scale [25]
There are 12 items in 5 subscales, including “Medical compliance,” “Collaboration with healthcare providers,” “Willingness to access and use information about medication,” and “Acceptance to take medication and how taking medication fits patient’s lifestyle”; each has 3 items. The scores range from 1 point for “Never” to 5 points for “Always,” including two reverse items, with a higher score indicating greater medication adherence.
Data Analyses
All analyses were performed with SPSS version 24.0 for Windows (Japan IBM, Tokyo) and SPSS AMOS ver 26.0 (Japan IBM, Tokyo). The analyses included item analysis, exploratory factor analysis (maximum-likelihood estimation, promax rotation, factor loading of 0.45 or more), reliability (Cronbach’s α coefficient, intra-class correlation coefficient [ICC] by re-testing), coexistense validity (Pearson’s correlation coefficient between the OSSA, SCAQ and medication adherence scale), multitrait scaling analysis, confirmatory factor analysis (analysis of covariance), and probability of use (Pearson’s correlation coefficient) by testing the hypothesis—“Patients with high opioid self-management scale scores have a low intensity of pain.”