Study Design and Participants
A quantitative survey of Vietnamese migrants was conducted from January 16th to March 16th, 2021. We selected the study participants from Kanto, Kansai, Tokai, and Kyushu, where Vietnamese migrants are highly geographically concentrated, according to the immigration data. A link to the questionnaire, along with an explanation, was posted to the Facebook groups of these communities. The questionnaire was created based on discussions with the leaders of the Vietnamese communities, and translated into Vietnamese by a member of the research group, who is a Vietnamese native (T.T.Hue), and checked by another Vietnamese research assistant. We conducted a pre-test among a sample group of 10 people, to confirm that the questions were comprehensible to the participants and relevant to the study.
All participants who clicked the link for the survey were screened for eligibility; the inclusion criteria included: being Vietnamese, aged 18 years or over, cumulatively living in Japan for 3 months or more. All those who completed the survey were provided an online-shopping voucher worth 1000 Japanese Yen.
In total, 600 participants answered the questionnaire, however, for the purpose of analysis, we excluded respondents whose status of residence could not be determined (n = 11, remaining 589 individuals). Most participants who completed the survey came from Kanto (241 people, 40.9%) and Kyushu (161 people, 27.3%) regions. The geographical distribution represented in the sample was somewhat similar to that recorded in the immigration data, although the sample overrepresented students and women (52.1% and 60.1%, respectively). Due to the lack of additional data on the general immigrant population in Japan, no further cohort comparisons can be made. However, in terms of the age profile of the participants, 85.2% of respondents were aged between 18–29 years old and 14.5% were aged between 30–39 years. Once again, this is broadly in line with immigration data, which indicates that these subgroups accounted for 78% and 18% of the Vietnamese migrants, respectively, in 2019.
Measurements
Sociodemographic factors
Sociodemographic factors included age, gender, marital status, educational background, length of stay in Japan, employment status, Japanese language skills, and health-related characteristics of the participants. According to marital status, participants were categorized into two groups: single or married; as per educational background, into two groups: high school or lower, university or graduate school; according to the length of stay, into three groups: less than 5 years, 5–10 years, above 10 years; and based on employment status into three groups: full-time, part-time, not working.
We assessed the participants’ Japanese language skills using a 7-item questionnaire, which evaluated their proficiency in conversation, reading and writing. Each item score ranged from 0 (not at all) to 3 (excellent). The total scores ranged from 0 to 21, with higher scores indicating better proficiency. These items have been previously used in another study of migrants in Japan(6, 25).
Moreover, we also asked questions about their health insurance, the need for Japanese language interpreters while visiting health facilities, and self-rated health status. We measured the participants’ self-evaluated health status using a single-item question with five possible responses (excellent, very good, good, fair, and poor). The responses were then categorized into two groups: (1) fair and poor, (2) excellent, very good and good, and treated as dummy variables with the omitted category being “excellent, very good and good”. This item has been commonly used in previous studies examining health status (26, 27), and migrant studies in Asia(28).
Pandemic-related risk factors
A series of 7 risk factors, likely to be associated with lower mental health, was included to assess the migrants’ experiences during the pandemic. These factors included the direct impact of the pandemic on health and healthcare access (COVID-19 infection history, difficulties in healthcare access), employment, and economic well-being (job loss and reduction in working hours, food insecurity, inability to return to home country). Regarding the COVID-19 infection history, we aggregated those with self-reported positive diagnostic cases and those who had lived with a confirmed case. For those who reported having contracted COVID-19, we asked the level of symptoms noted, based on the guidelines issued by the Ministry of Health, Labor and Welfare of Japan, i.e., (1) No symptoms, (2) Mild symptoms, (3) Moderate level 1, (4) Moderate level 2, (5) Serious symptoms. We also included items regarding healthcare access, i.e., whether the pandemic had affected access to usual medical care.
Regarding the employment and economic impact, we asked the participants whether they had lost their job or faced reduced working hours, and categorized the responses into two groups: Yes (1 = job insecurity), No (0 = no job insecurity). Furthermore, we defined food insecurity for those who had to reduce meal sizes or skip meals during the pandemic. We also asked the participants whether they tried to return to their home country and aggregated those who reported that they could not return due to lack of money or travel restrictions. These items were adapted from (29–31).
Protective factors
We screened the respondents for self-perceived social support using the Multidimensional Scale for Perceived Social Support (MSPSS). The MSPSS consists of 12 items, reflecting support from family, friends, and significant others, with each domain containing four items, scored as 1 (very strongly disagree) to 7 (very strongly agree). The total scores range from 12 to 84, with higher scores indicating greater perceived social support. This tool was used in the study of (32) on the migrant population in China.
Mental health outcome
We assessed the presence of depression during the last two weeks using the Center for Epidemiologic Studies Depression Scale (CES-D) Scale, a common screening tool for mental disorder(33–35). The reliability and construct validity of this instrument have been reported in multiple investigations in Japan (36) and Vietnam (37). The CES-D consists of 20 items for depression and is scored as 0 (rarely or none of the time) to 3 (mostly or all the time). The total scores range from 0 to 60, and we defined scores of 16 or more as having depression. The participants were divided into two groups: those with depressive symptoms (CES-D score > = 16 points), and those with no depressive symptoms (CES-D < 16 points).
Statistical Analysis
Descriptive statistics were used to summarize sociodemographic characteristics and characterize COVID-19 related experiences using percentages for categorical variables, and mean and standard deviation (SD) for continuous data. Using established cut-off scores of the CES-D, we examined the frequency (percentage) of participants who screened positive for possible cases of depression. We compared the different characteristics and the differences in CES-D scores among the resident status groups using Pearson’s chi-square test for categorical variables and one-way ANOVA for continuous variables.
We employed a multivariable logistic regression model with three categories of explanatory variables, i.e., sociodemographic characteristics, pandemic-related risk factors, and protective factors, to identify the factors associated with depression. Additionally, the associations of these factors for different resident status groups were examined separately. As the permanent resident group was too small for meaningful comparison in multivariable testing, only students’ and technical trainees’ data were analyzed. The adjusted odds ratios (AOR) and 95% confidence intervals (CIs) were used as measures of association. We analyzed the data using STATA Statistics version 13. No multicollinearity was found among the predictor variables. A p-value less than 0.05 indicated statistical significance.
Ethics Approval
We received ethics approval from Kyorin University Research Ethics Committee (No. 43). The participants were invited to participate in the survey via a website developed by us, which included the study’s information sheet and consent form. A consent form was obtained from all participants prior to the web-based survey.