Participants and procedure
The survey company (Macromill, Ltd.) recruited potential participants from a pool of 10 million who were registered in a national database and sent questionnaires to them online over 5 days in April 2017. Eligible participants were those with cancer followed in an outpatient clinic and aged 16-39 years old at the time of survey. The 15-39 age range is the standard used by the National Cancer Institute in the United States. According to the Ethical Guidelines for Medical and Health Research Involving Human Subjects formulated by the Ministry of Health, Labor and Welfare of Japan, minors who have graduated from junior high school or are 16 years of age or older can be judged to have sufficient judgment ability regarding the conduct of the study. For these reasons, we set the eligibility age from 16 to 39 years. Potential participants first read introductory statements that summarized the contents of the questionnaire and explained they could withdraw at any time if they wished so. Responses were considered consent to participate. Because the legal adult age is 20 years old in Japan, those who were 16-19 years old were asked to confirm the consent of their parents. Responses to the questionnaire were voluntary, and confidentiality was maintained throughout all investigations and analyses. This study was approved by the Institutional Review Board and Ethics Committee of the National Center for Neurology and Psychiatry, Japan, and was conducted in accordance with the principles laid down in the Helsinki Declaration.
Measures
Supportive care needs: The short-form Supportive Care Needs Survey questionnaire (SCNS-SF34)
The SCNS-SF34 is a 34-item self-report questionnaire. It measures perceived needs of cancer patients in five domains: psychological (10 items), health system and information (11 items), physical and daily living (5 items), patient care and support (5 items) and sexuality (3 items). A 5-point Likert scale is used where 1 corresponds with no need (not applicable), 2 no need (satisfied) and 3, 4 and 5 with low, moderate and high levels of unmet need, respectively. Subscale scores were obtained by summing the individual items. In addition, the total score was obtained by summing all the subscales (range = 34-170). A higher score indicated a higher perceived need. As an alternative use, the scale can be used to obtain information on the presence/absence and number of perceived unmet needs (a rating of 3 or higher was regarded as an unmet need), depending on the researcher’s clinical question. Participants were instructed to circle the number that best describes whether they have needed help with each of the statements in the past month. The validity and reliability of the Japanese version of the SCNS-SF34 have been established [18].
Perceived social support: The short-version Multidimensional Scale of Perceived Social Support
It measures social support from family, friends and significant other with seven items. A 7-point Likert scale is used, and a low score for a scale represents poor social support. The validity and reliability of the Japanese version of the short-version Multidimensional Scale of Perceived Social Support have been confirmed [19].
Psychological distress: The Kessler Psychological Distress Scale (K6)
The K6 scale is a self-rated six-item questionnaire exploring the frequency of psychological distress during past 30 days. Responses are rated on a five-point scale ranging from “none of the time” (0) to “All the time” (4); scores range from 0 to 24. High scores indicate more severe mental disorders. The validity and reliability of the Japanese version of the K6 have been established [20].
Quality of life (QOL): The EuroQol 5 Dimensions 5 Levels (EQ-5D-5L)
The EQ-5D-5L measures five dimensions; mobility, self-care, usual activities, pain/discomfort, and anxiety/depression. A 5-point Likert scale is used and a high score for a scale represents a high level of symptomatology and problems. The validity and reliability of the Japanese version of the EQ-5D-5L have been confirmed [21].
Demographic and medical characteristics
The participants were asked their demographic and clinical information, as listed in Table 1. Age, sex and occupational status were obtained from the basic information of the monitoring system.
Sample size calculations
We planned to perform a multiple regression analysis to examine the factors related to the supportive care needs, and we calculated that 10 times as many subjects as the number of independent variables would be required. Considering the missing data, the planned number of subjects was 150.
Statistical analysis
First, we described the prevalence of unmet needs. Second, to identify potential demographic, biomedical, and psychosocial factors associated with a high degree of unmet needs, we conducted a preliminary univariate analysis. In this preliminary analysis, the presence or absence of unmet needs was entered as dependent variables. The presence of unmet needs means that there is more than one item rating of 3 or higher in the SCNS-SF34. Demographic and clinical variables and social support were entered as independent variables. For the univariate analyses, an unpaired t-test, Mann-Whitney test and chi-square test were conducted, as appropriate. After the univariate analyses, we used a multiple regression analysis to examine the final factors associated with patients’ unmet needs. The dependent variables were total score (total needs) and the score of each domain. Independent variables with a p-value of less than 0.05 in the preliminary univariate analyses were entered into the multiple regression analysis. We entered associated factors (sex, age, age of first diagnose, PS, marital status and chemotherapy) which were reported in previous research [4, 7-9, 11, 12] as independent variables as well. A backward stepwise selection method was used to reduce non-significant variables from the models. To investigate the association between the patients’ supportive care needs and psychological distress and QOL, Pearson’s correlation analysis was conducted. Data were analyzed with the SPSS version 25.0 (IBM). All of the tests were two-tailed, with a p-value of <0.05.