Mental health status and quality of life among Cambodian migrant workers in Thailand

Study design and sample

A cross-sectional study was conducted by using a multistage random sampling method to choose the study sample. Data were collected between March 2018 and May 2018. Cambodian migrant workers are required to register with the Department of Employment Office in each province in order to benefit from Universal Health Coverage (UHC), and to be bestowed with legal working status in Thailand. We applied for and gained permission to use this list from the Department of Employment Office in Thailand. For Sa Kaeo and Surin provinces, a total of 24,256 Cambodian migrant workers were listed during time of study. Finally 1,211 sample size was calculated following the formula from Hsieh FY et al.. (1998) as follows:

The approximate sample size was 435 which were further adjusted to control the over-fitting using the rho (ρ) of 0.65 and variance inflation factor (VIF) equal to 2.85. Therefore, the total number of the sample was 1,211. The sample size was calculated to be 1,211. The sampling process was achieved as follows: (I) Sa Kaeo and Surin provinces were randomly selected from the total number of 7 eastern and northeastern Thai provinces that share a boarder with Cambodia, representing a total migrant population in the two provinces of 24,256 (Sa Kaeo represents 89% of this with 21,619 migrant workers, and Surin the remaining 11%, with 2,637 migrant workers). (II) The sample size was split by proportion to the relative populations of the two provinces, with 1,053 samples allocated to Sa Kaeo province and 158 to Surin. (III) Simple random sampling was used to select 8 out of 26 districts. (IV) Systematic random selections were made from the name lists of the total migrant worker population in each of these districts until the completed sample size was achieved.

Inclusion criteria: All participants in the study were 18 years of age or above, and had been working in Thailand for at least 1 month before the interview period. Having the required legal documents including a passport, work-permit, border pass required for legal migrant workers, and good physical health and mental health and were willing to participate in the study.

Exclusion criteria: Any person with a disability was excluded. Participants with unofficial documents and who were not willing to participate in the study were also excluded.

The data was collected through in-person interviews conducted by trained interviewers. Interviews took place between March 2018 and May 2018. Prior to data collection five research assistants were trained one-day training on the study objectives and how to administer the questionnaires. Once they understood the data collection process research assistants were paired to test the questionnaire on each other to further ensure they were familiar with all parts of the questionnaire.

All participants were informed about the purpose, benefits and assured of confidentiality before signing the consent for the study. For convenience, researcher also asked permission from the employer of the migrant workers to interview participants at time that would minimize disruption to their working hours. If participants were illiterate researchers asked a literate volunteer to witness the accurate reading of the consent form, allow the participant to ask any questions and then subsequently signed on their behalf. This study was approved by the Khon Kaen University (KKU) committee for research ethics in human research (Reference no. HE602361), Khon Kaen, Thailand.

Research instruments

A structured questionnaire was developed from reviewed literatures based on research questions, first in English and was then translated into Khmer using forward and backward translation procedures. The questionnaire consisted of 4 sections which were a) individual characteristics and socio-demographic factors, namely gender, age, marital status, educational attainment, occupation, financial status, work environment, incidence of work injury, residential arrangement, house tenure, distance from house to community center (km), daily travel to work, the incidence of work-related diseases during past 12 months, and smoking status. A copy of the questions asked are provided as extended data, b) the Perceived of Stress Scale (PSS) of Cohen et al. (1983), d) depression Scale (CES-D) of Radloff et al. (1977) with Cronbach’s alpha of 0.83, and c) WHOQOL-BREF Khmer version with Cronbach’s alpha of 0.77. The questionnaire was undergone content validation by five experts and was revised to improve validity.

Statistical analysis

Demographic characteristics of the participants were described using frequency and percentage for categorical data and using mean and standard deviation for continuous data. Inferential statistics comprising simple logistic regression bivariate and multivariate models were used in a multilevel mixed-effects model to reduce clustering effects. Confidence intervals (CI) were taken at 95% and statistical significance was considered at p<0.05. All analyses were performed using Stata version 10.0 (Stata Corp, College Station, TX).

Demographic characteristics

Of the 1,211 Cambodian migrant workers, 50.37% were male and 50.29% were working in the agricultural sector. The mean age was 32.54 (±11.13) years (range: 18–67) (Table 1). The majority of respondents were married (62.59%). Most had no formal education (42.69%). The median monthly personal income was 7,500 (range: 7,500-20,000) Baht (equivalent to 237 USD at time of publication; Table 1). Mental health problems were common, with 57.72% of participants reporting moderate levels of perceived stress (32.04% low level and 10.24% high level). Many respondents reported symptoms of depression (69.69% compared to 30.31% not reporting these symptoms) (Table 2). Most of the participants (55.33%) had a moderate QOL level, 34.52% had low QOL and 10.16% had high QOL (Table 3).

The bivariate analysis using simple logistic regression (SLR) indicated that having a high perceived stress score and having depression symptoms were significantly associated with poor QOL (p<0.001 and p=0.001 respectively). However, sex, age, marital status, educational attainment, job category, financial status, working environment, suffering a work injury in the past 12 months, residential arrangement, house tenure, distance between house and community center, daily travel to work, suffering work-related diseases in the past 12 months, and smoking were also all significantly associated with poor QOL (Table 4).

The final model was constructed using multivariate analysis, controlling for clustering effects using multilevel logistic regression (MLR) and control covariates. This final model indicated that factors significantly associated with poor QOL were a high perceived stress level (AOR=3.64; 95%CI: 2.41-5.49; p<0.001); living with family/relatives (AOR=3.63; 95%CI: 2.42-5.45; p<0.001); living in housing provided by their employer (AOR=2.66; 95%CI: 1.74-4.08; p<0.001); commuting to work using their employer’s vehicle or by foot (AOR=1.64; 95%CI: 1.11-2.42; p=0.012); and living more than 1 km from the community center (AOR=1.41; 95%CI: 1.07-1.87; p=0.016). However, having symptoms of depression was not associated with poor QOL [Table 1–Table 5].

Explaining the findings

The prevalence of low QOL in Cambodian migrant workers was 34.52% (95%CI: 31.84-37.20) with moderate QOL at 55.33% (95%CI: 52.52-58.13) and high QOL at 10.16% (95%CI: 8.45-11.86). This result is in contradiction to a previous study on the QOL of Burmese migrant workers in the Chiang Rai province of Thailand¹⁰, which reported 0.20% as having low QOL, 56% moderate QOL and 43.8% high levels of QOL. These differences in the prevalence of low QOL may be due to the target population of the previous study being mainly women between 18 and 29 years old, mostly working in industry. The working and living environments of those in industry are less objectionable when compared to that of agricultural workers. In contrast, the population surveyed in the present study had a gender balance nearer parity, mostly within the age range 20 to 49 years, and half of the sample worked at agriculture. Moreover, the socio-cultural settings of Cambodian migrant workers and Burmese migrant workers are different. The findings of the present study are in agreement with a study conducted in Phang-Nga province of Thailand that also reported most migrant workers as having low to moderate QOL¹². A study in Dhaka city, Bangladesh showed that 94% of adult migrant workers in garment manufacture had low QOL, with only 3.25% and 2.75% having moderate and high QOL respectively. This may be due to the study in Bangladesh recruiting more female than male workers; within Indianite cultures, the female gender is devalued and females are perceived as the inferior gender, both physically and psychologically, resulting in very low QOL compared to males¹³. The above results demonstrate the inconsistency of QOL findings across the literature. This inconsistency likely results from the multicultural differences in context and setting of the living and working environments of migrant workers. Another example is the report of female domestic migrant workers in Singapore exhibiting a high QOL within three domains out of four, with the social relationship domain exhibiting a low score and also being associated with stress¹⁶ [Figure 1].

Figure 1. Forest plot of the adjusted odds ratios and 95%CI: confidence intervals of the association between common mental health problems and poor quality of life among Cambodian migrant workers in Thailand using multilevel logistic regression (MLR).

Moreover, mental health problems were reported among migrant workers. The prevalence of moderate to high perceived stress was 67.96% (95%CI: 65.33-70.59) and symptoms of depression was 69.69% (95%CI: 67.10-72.29). A comparable study in Europe²¹ reported 63% of low-skilled workers to exhibit symptoms of distress. Moreover, stress and depression are often cited as the most predominant factors associated with reduced QOL scores, with some studies finding depression to be the most influential on QOL, and vice-versa²². Additionally, stress is commonly reported to impact QOL for migrant workers; such as 62.2% of white-collar migrant workers in China reporting work related-stress²³.

After controlling for other covariate factors, the factors that were indicated as being significantly associated with poor QOL were high perceived stress (AOR=3.64; 95%CI: 2.41-5.49; p<0.001); living with family/relatives (AOR=3.63; 95%CI: 2.42-5.45; p<0.001); living in employer-provided housing (AOR=2.66; 95%CI: 1.74-4.08; p<0.001); commuting to work using their employer’s vehicle or by foot (AOR=1.64; 95%CI: 1.11-2.42; p=0.012); and living more than 1 km from the community center (AOR=1.41; 95%CI: 1.07-1.87; p=0.016). However, having symptoms of depression was not significantly associated with poor QOL in the final model.

Factors associated with perceived stress have a strong association with poor QOL (AOR=3.64; 95%CI: 2.41-5.49; p<0.001). In stressful situations, migrant workers may turn to alcohol and smoking. This in turn will increase the economic burden, impact on their capacity to work, may cause conflicts in the family, and can sometimes be associated with gambling. Moreover, stressful situations themselves may cause workers to lose focus on their job and worry about earning money to feed their family or reduce the debt incurred in the process of migration. Many workers worry about the expiration of work permits and health insurance documents. A study in China has shown that work related stress was associated with poor QOL^15,23. A previous study among migrant workers in India found 14.6% workers to have poor QOL and 25.5% to be in psychological distress, with QOL being significantly (p≤0.05) associated with the factors of age, marital status, low education status²⁴. Stress was associated with both depression and anxiety study among Burmese migrant workers on the borders²⁵. Lifestyle factors were also found to impact migrant workers in the study, for example smoking was associated with low QOL. Smoking was common among the male migrant workers surveyed in the present study in terms of the habitual lifestyle associated with a stressful working environment. Smoking severely increases the economic and health burden on migrant workers. In the study rural to urban Chinese migrant workers, it was found that unhealthy lifestyles caused by many factors, such as the kind of job that they were doing, working in too small a space, low income and long working hours, were significantly related to the poor mental health of migrant workers²⁶.

Living with family and relatives (AOR=3.63; 95%CI: 2.42-5.45; p<0.001) means that migrant workers have greater levels of responsibility in their daily life. In a Bangladeshi study, living with at least one family member was highly associated with low QOL (p=0.01)¹³. In the context of Cambodia, migrant workers often live in a small cottage where they work together with many family members, relatives or friends. If this living space becomes too crowded, workers will have difficulty sleeping and further decreased QOL.

Having their living accommodation provided by their employer (AOR=2.66; 95%CI: 1.74-4.08; p<0.001) was more likely to be associated with low QOL than living in rented housing. In the present study, many migrant workers who lived inside farmworker camps or construction site camps were reported to live in employer-owned accommodation, and also were more likely to be exposed to hazardous living environments²⁷. There are many potential factors associated with this. For example, in worker camps, even though most employer provide employees with a living place free of charge, migrant workers have to build the cottage themselves, often resulting in poor build-quality in terms of dirt, dust and a lack of mains electricity supply. Living together with strangers is not good for psychological health, and often migrant workers from Cambodia end up cohabiting with many colleagues from different places. However, renting a house may cause a low QOL depending on the study context¹³. Migrant workers in China showed that residential satisfaction was at a moderate level²⁸.

Commuting daily to the workplace using their employer’s vehicle or on foot (AOR=1.64; 95%CI: 1.11-2.42; p=0.012) was more highly associated with low QOL compared to travelling to work using their own vehicle. Daily basic transportation was important among migrant workers where most were living in isolated rural areas. A study among Burmese migrant workers in Thailand found transportation difficulties to be barriers to accessing health services²⁹.

Living further than 1 km away from the community center was associated (AOR=1.41; 95%CI: 1.07-1.87; p=0.016) with low QOL. It might be that living far from a community center makes it difficult to buy food or to access essential services. Migrant workers who did not own vehicles in Thailand, like bicycles or motorbikes, were more likely to have a low quality of life. This may be further confounded by an unfamiliarity with the area where they live and work, making it difficult to locate the community center if they live far from it. This will make it more difficult when they want to buy food at the market or access to other services in Thailand. Similarly, a study of workers in Russia showed that distance from house to workplace had a significant impact on job satisfaction³⁰.

Underlying data

Figshare: 1. Dataset 1 Raw data from a survey mental health status and quality of life among Cambodian migrant workers in Thailand.xlsx. https://doi.org/10.6084/m9.figshare.12769856.v2

Extended data

Figshare: 1. Dataset 1 Raw data from a survey mental health status and quality of life among Cambodian migrant workers in Thailand.xlsx. https://doi.org/10.6084/m9.figshare.12769856.v2

This project contains the following extended data:

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).