UTILIZATION OF HEALTH CARE SERVICES AMONG INTERNAL MIGRANTS IN HANOI AND ITS CORRELATION WITH HEALTH INSURANCE: A CROSS-SECTIONAL STUDY

Abstract

Background

Economic transition (DoiMoi) in the 1980s in Viet Nam has led to internal migration, particularly rural-to-urban migration. Many studies suggested that there is a difference between non-migrants and migrants in using health care services. Current studies have mostly focused on migrants working in industrial zones (IZs) but migrants working in private small enterprises (PSEs) and seasonal migrants seem to be ignored. However, these two groups of migrants are more vulnerable in health care access than others because they usually work without labor contracts and have no health insurance. The study aims to compare the utilization of health care services and explore its correlated factors among these three groups.

Methods

This cross-sectional study included 1800 non-migrants and migrants aged 18–55 who were selected through stratified sampling in Long Bien and Ba Dinh districts, Hanoi. These study sites consist of large industrial zones and many slums where most seasonal migrants live in. A structured questionnaire was used to collect information on health service utilization in the last 6 months before the study. Utilization of heath care services was identified as “an ill person who goes to health care centers to seek any treatment (i.e. both private and public health care centers)”.

Results

644 of 1800 participants reported having a health problem in the last 6 months before the study. Among these 644 people, 335 people used health care services. The percentage of non-migrants using health care service was the highest (67.6%), followed by migrants working in IZ (53.7%), migrants working in PSE (44%), and seasonal migrants (42%). Multivariate logistic regression showed migrants, especially seasonal migrants and migrants working in PSE, were less likely to use health care services (OR=0.35, p=0.016 and 0.38, p= 0.004, respectively), compared to non-migrants. The study also found that having no health insurance was a risk factor of the utilization (OR=0.29, p<0.001). Other factors such as gender, age, marital status, socioeconomic status, and monthly income were not related to the utilization of health care services.

Conclusion

Seasonal migrants have the worst utilization of health care services, followed by migrants working in PSE, migrants working in IZ, and non-migrants. Health insurance is an important factor relating to the utilization. Accordingly, health insurance coverage needs to be increased if utilization of health care services for the whole population, particularly migrant population, is to be improved.

BACKGROUND

Economic transition (DoiMoi - Renovation) in the 1980s in Viet Nam has led to internal migration, particularly rural-to-urban migration^1,2. Migrants are generally healthy because they need to adapt to working requirements, their living environment, and the mobility of their place of residence^3,4. However, their health quality declines over time as a result of many pressures of life, especially financial pressure^5,6. In addition, migrants are more susceptible and vulnerable to ill-health effects than non-migrants⁷. Therefore, one can assume that their health would often put them at higher need of health care services than non-migrants. Nevertheless, evidence showed that migrants seemed to have less opportunities of using health care services than non- migrants^8,9.

In Viet Nam, previous studies stated that migration did not brought additional pressures on the health system at the destination. In other words, the available health system could meet the increase of migration flows^8,10. However, some studies stated a difference between non-migrants and migrants in using health care services in Viet Nam. The 2004 Migration Survey showed that the proportion of medical service utilization in migration population was lower than that in non-migration population⁸. Other studies also identified that one of the most popular reasons that affect the utilization of health care services was medical fees that were usually beyond what migrants could afford^11,12. Additionally, a number of studies found that there was a relationship between the level of health insurance coverage and utilization of health care services. Zuvekas and Taliaferro (2003) stated that increasing health insurance coverage would undoubtedly increase access for all people in general and reduce racial disparities^13,14. Peng et al (2010) also identified migrants without health insurance are much less likely to use health care services than migrants with health insurance¹².

Health insurance of migrant population in Viet Nam remains at a low level of coverage. The 2004 Migration Survey found that only 31% and 42% migrant workers in the Northeast and Southeast Industrial Zone, respectively, have compulsory health insurance that is mostly paid by employers; meanwhile the proportion of health insurance for non-migrants was around 50%⁸. Migrants, who are not suitable for compulsory health insurance (e.g. migrants without labor contract), can join voluntary health insurance and pay monthly insurance membership fees themselves¹⁵. However, migrants usually ignore voluntary health insurance because they do not want to use any part of their low income for health insurance⁹. All these disadvantages might limit migrant usage of health care services at destination.

The economic transition has also led to considerable reforms in the Vietnamese health system. Due to the transition, the health system was subsidized by the Government and provided free health care services to the whole population. During the transition period, several reforms were implemented that aimed to realize the goal “the government and people working together”. Such reforms have resulted from health policies on charging partial user fees, private health practice in 1989, health insurance in 1992, and reduction and exemption of user fees for the poor, minorities, and poor regions/areas in 1994¹⁶. This has brought not only numerous improvements to health care system, but also several opportunities for health care access to population. However, out-of-pocket household health expenditure in fact has still accounted for a large proportion of total health expenditure¹⁶. For instance, the proportion in 2004, 2005 and 2006 were 63.9%, 65.8% and 62.8%, respectively. This has had a negative impact on the goals of equity and efficiency in health care. The payment of direct health care fees for an inpatient among the poor is around 17 months of non-food household expenditure per capita while the payment among the rich group accounts for 8 months. Thus, the poor and other vulnerable groups (e.g., migrants, patients of chronic diseases, low-income population) prefer low-price health care services that are usually low quality or no treatment.

Most recently, the Vietnamese Ministry of Health has just approved a new cost norm for health services user fees that is much higher than the previous cost norm¹⁷. There has been a concern that the new cost norm may significantly affect the poor and other vulnerable groups because they would have a lower chance to access quality services with affordable prices.

Current studies regarding the utilization of health services in migrant populations in Viet Nam have mostly focused on migrants working in industrial zones (IZ) because this population accounts for the highest proportion of rural-to-urban migrant population^18,19. However, there are still many different sub-groups of rural-to-urban migrants such as migrants working in private small enterprises (PSE), migrants working in construction projects, truck drivers, seasonal migrants and so on^9,20,21. Among these sub-groups, two groups of migrants working in PSE and seasonal migrants also account for the second and the third proportions of rural-to-urban migrants. In addition, migrants of these two groups usually work without labor contracts in private enterprises or self-employed; thus, they might be less likely to have health insurance⁹. Therefore, their utilization of health care services could be limited.

The paper aims to compare the utilization of health care services among four migration groups: non-migrants, migrants working in industrial zones, migrants working in private small enterprise, and seasonal migrants. It also aims to explore correlated factors of health care services utilization among the migrant populations.

METHODS

Study sites and participants

The study was conducted in Hanoi, Viet Nam. Long Bien and Ba Dinh districts were selected as two study sites because these districts include several rural-to-urban migrants.

A non-migrant is defined as a person who has had permanent residence (ho khau – in Vietnamese) and lived at the study site for at least 5 years prior to the time of the study. Meanwhile, migrants (i.e. migrants working in IZ and migrants working in PSE) are defined as a person who moves from another province to Hanoi and stays there from 6 months – 5 years. The definition is exactly the same as defined by the 2004 Viet Nam Migration Survey and the 2009 Censuses^19,22.

However, the definition does not capture seasonal migrants who leave their home town to another town for a short time (i.e. 4.6 months in average). Hence, Brauw and Harigaya defined these seasonal migrants as “members of the household who left for a part of the year to work, but are still considered household members”²³. In addition, Duong and Liem have defined temporary/seasonal migrants as “those who came from a rural area within 6 months or less prior to the time of the interview and who do not have a permanent household registration in the city of destination”⁹. Therefore, a seasonal migrant in this paper is defined as a person who leaves his or her home town to Hanoi for less than 6 months and gets a temporal job without a labor contract (e.g. porters, street vendors, waiters, maids, and other services and so on).

All participants in the study were 18 – 55 years old, which reflects the working age in Viet Nam (18 – 60 for males and 18 – 55 for females). Participants were randomly collected from the sampling frame of each group. The sampling frame of non-migrants was based on the household registration (ho khau) while the sampling frame of migrants came from temporary registration (tam tru) and other non-registered migrants identified by heads of resident groups (to dan pho).

Study design and sample size

The cross-sectional study aimed to compare two proportions of health care utilization between migrants and non-migrants with specified relative precision. We estimated two parameters p₁= 0.25 and p₂= 0.35 that were taken from proportions of access to health care services in a pilot study. Thus, the study included 450 participants in each group (i.e. 1800 participants in the total). In practice, 1900 participants were selected to ensure sufficient sample size in the case of non-response or missing information. As a result, 1826 participants were interviewed (i.e. response rate was 96%); however, 26 participants were missed important information (e.g. gender, occupation, utilization of health services, and health insurance). Therefore, 1800 participants were used in data analysis.

Questionnaire and terminology

Information on access to health care and its correlates were collected using structured questionnaire, which was piloted in both non-migrant and migrant populations. It included (i) background information such as age, education, occupation, marital status, monthly income and expenditure, and working time; (ii) living conditions of participants such as house status, water supply, toilet, and durable living assets; (iii) health care insurance; and (iv) health problem and utilization of health care services in the last 6 months prior to the study. The questionnaire was validated by a previous qualitative study that published by Anh et al. (2011)²⁴.

The interviews were conducted face-to-face between interviewers and interviewees at interviewees’ places of residence during their free time. The interviewers in this study were Master’s students at the Hanoi School of Public Health and were trained by principal investigators. Informed consent (i.e. written informed consent) was obtained from each participant at the start of the interview and this study was approved by the Institutional Review Board of the Hanoi School of Public Health and the Ethikkommissionbeider Basel (EKBB) in Basel, Switzerland.

In the study, a health problem was identified as “a person whose illness prevents regular work for 1 day”^8,9. Utilization of heath care services was identified as “an ill person who goes to health care centers to seek any treatment (i.e. both in private and public health care centers)”. Meanwhile, ill persons who do nothing or buy medicine at a pharmacy without a doctor’s prescription or use medicine by themselves were identified as “no utilization of health care services”^8,9.

Terminology of living conditions in the study was based on the 2009 National Census and the 2004 Migration Survey^18,25, which categorized housing status, for instance, into three levels: permanent, semi-permanent, and simple depending on the main construction materials of the pier, the roof, and the outer. The number of “simple” levels is small because this level in fact is not an issue in urban settings. Therefore, this study used two levels of housing status: permanent and non-permanent. Through conducting principal components analysis (PCA) of living assets, socio-economic status (SES) was categorized into three different levels including high, average, and low²⁶.

Working time was defined as two groups: working in regular time or giohanhchinh -in Vietnamese (i.e. 7.30 am – 5:30 pm) or in shifts and flexible working time or nghetu do – in Vietnamese (i.e. self-employed). This classification is suitable to characteristics of working time for most non- and migrants in Vietnam. In addition, this study used 3 millions VND as a cutoff-point for monthly income of each participant to categorize into two groups: less than 3 millions VND and 3 millions VND and above. The classification of working time and monthly income is in line with other studies on migrants in Vietnam^6,22.

Analysis methods

Prior to the analysis, all questionnaires were reviewed for accuracy. Ten percent of data was independently entered by two research assistants to check data entry procedure. All data was stored in EpiData 3.1 and transferred to Stata 10.0 for analysis. Tabular technique and chi-square were used to compare characteristics among non-migrant and migrant populations. Multiple logistic regressions were then used to identify correlates of utilization of health care services. Assumptions of models were evaluated before producing the final models and these models were also tested by using the Hosmer-Lemeshow test. All tests in the study used a significance level of 0.05.

RESULTS

Characteristics of study sample

The results of Table 10.1 show different characteristics among migrant populations and between non-migrant and migrant populations. The proportion of females in seasonal migrants is higher than the proportion in other migrant groups (74.2% compared to 44% of migrants working in PSE and 50.2% of migrants working in IZ). Seasonal migrants are also older than migrants working in PSE and migrants working in IZ. In addition, working time, monthly income, and living condition are significantly different among migrant populations. Most migrants working in IZ (about 93%) work in regular time or in shifts, while most seasonal migrants (around 99%) are self-employed including street vendors, porters, and motorcycle taxis. Monthly income of seasonal migrants is the lowest compared to non-migrant and other migrant populations. Additionally, living condition/SES of seasonal migrants is also the worst.

Table 1 also shows the coverage of health insurance among non-migrant and migrant populations. The proportion of migrants working in IZ is the highest, compared to the two other migrant groups and even to non-migrants (e.g. 77.78% compared to 26.22% for migrants working in PSE, 22.89% for seasonal migrants, and 55.56% for non-migrants).

Table 1.

Characteristics of study samples

Characteristics	Non migrants		Migrants in IZ		Migrants in PSE		Seasonal migrants		Total
	n	%	n	%	n	%	n	%	n	%
	450	100	450	100	450	100	450	100	1800	100
Gender
Male	163	36.2	224	49.8	252	56	116	25.8	755	41.9
Female	287	63.8	226	50.2	198	44	334	74.2	1,045	58.1
Age groups
< = 30	176	39.1	406	90.2	265	58.9	131	29.1	978	54.3
31–40	151	33.6	37	8.2	104	23.1	152	33.8	444	24.7
>40	123	27.3	7	1.6	81	18	167	37.1	378	21.0
Marriage
Single	85	18.9	309	68.7	171	38	43	9.6	608	33.8
Married	365	81.1	141	31.3	279	62	407	90.4	1,192	66.2
Working time
Regular time/in shift	214	47.6	418	92.9	207	46	6	1.3	845	46.9
Self-employed	236	52.4	32	7.1	243	54	444	98.7	955	53.1
Monthly income
< 3 million VND	116	34.7	114	26.9	111	28.5	197	45.9	538	34.2
≥ 3 million VND	218	65.3	309	73.1	278	71.5	232	54.1	1037	65.8
Living assets/SES
High	420	93.3	126	28	162	36	13	2.9	721	40.1
Average	14	3.1	151	33.6	103	22.9	92	20.4	360	20
Low	16	3.6	173	38.4	185	41.1	345	76.7	719	39.9
Housing status
Permanent	417	92.7	271	60.2	177	39.3	62	13.8	927	51.5
Non-permanent	33	7.3	179	39.8	273	60.7	388	86.2	873	48.5
Health insurance
Yes	250	55.6	341	75.8	118	26.2	103	22.9	812	45.1
No	200	44.4	109	24.2	332	73.8	347	77.1	988	54.9

Note: All characteristics in the table are significantly different among populations (p<0.001)

Health service utilization

In this study, there are a total of 644 participants (i.e. 177 non-migrants, 163 migrants working in IZ, 120 migrants working in PSE, and 183 seasonal migrants) having a heath problem in the last 6 months prior to the study.

However, only 335 of these 644 participants used health care services. The percentage of non-migrants using health care service is the highest 71.19%, 95% CI: 64.45–77.92%), followed by migrants working in IZ (53.37%, 95% CI: 45.63–61.11%), seasonal migrants (40.98%, 95% CI: 33.79–48.18%) and migrants working in PSE (38.84%, 95% CI: 30.03–47.65%).

Generally, utilization of health care services is significantly different between participants with and without health insurance (p<0.001). Indeed, the proportion of health care services utilization is about 66% for people who have health insurance while it is about 41% for people who have no health insurance (Table 2). Similarly, Table 2 also shows that the utilization is significantly different among education levels, housing status, and SES groups (i.e. living assets).

Table 2.

Utilization of health services in the last 6 months across different correlates

	Utilization of health care facilities	Total	p	95% CI of p	Chi-square p-values
	Yes	n
Health care insurance					<0.001
Yes	185	278	66.55	60.97–72.13
No	150	366	40.98	35.92–46.05
Education					0.02
Secondary and less	127	271	46.86	40.88–52.84
High school	126	240	52.5	46.14–58.86
Colleges and above	82	133	61.65	53.28–70.03
Housing status					<0.001
Non-permanent	199	325	61.23	55.91–66.56
Permanent	136	319	42.63	37.18–48.09
Gender					0.196
Male	99	205	48.29	41.39–55.19
Female	236	439	53.76	49.08–58.44
Age groups					0.881
≤30	175	334	52.4	47.01–57.78
31–40	81	153	52.94	44.94–60.94
>40	79	157	50.32	42.41–58.23
Religion					0.46
Non	323	620	52.1	48.15–56.04
Others	12	24	50	28.43–71.57
Marriage status					0.073
Single	91	195	46.67	39.6–53.73
Married	244	449	54.34	49.72–58.97
Living assets					<0.001
Sufficient	168	270	62.22	56.4–68.04
Normal	61	121	50.41	41.38–59.45
Insufficient	106	253	41.9	35.78–48.01
Working time					0.073
Regular time/in shifts	163	292	55.82	50.09–61.55
Flexible	171	351	48.72	43.46–53.97

The results above suggest possible correlates of health care utilization. Table 3 is logistic regression model for identifying these correlates. As a result, there is significant difference of health care utilization among non-migrant and migrant populations. Seasonal migrants are people who use health care services the least, followed by migrants working in PSE and migrants working in IZ and non-migrants (OR=0.35, 0.38, and 0.55 compared to non-migrants, respectively) (Table 3). Moreover, table 3 also shows the significant relation between health service utilization and health insurance coverage. Persons who have no health insurance are much less likely to use health care services compared with people who have health insurance (OR=0.29, p<0.001). Other factors such as gender, age, marriage status, SES, working time and monthly income are not related to the utilization of health care services.

Table 3.

Logistics model of correlated factors for health service utilization

		OR	95% CI		Wald p-values
Population	Non-migrants	1	–	–	–
	Migrants working in IZ	0.55	0.27	1.16	0.116
	Migrants working in PSE	0.38	0.19	0.74	0.004
	Seasonal migrants	0.35	0.15	0.82	0.016
Gender	Male	1	–	–	–
	Female	1.19	0.78	1.80	0.425
Age	≤30	1	–	–	–
	31–40	1.14	0.68	1.91	0.61
	>40	1.04	0.60	1.80	0.892
Education	Secondary school and less	1	–	–	–
	High school	0.95	0.57	1.58	0.846
	College and above	1.20	0.62	2.30	0.591
	Single	1	–	–	–
Marital status	Married	1.36	0.80	2.31	0.249
	Permanent	1	–	–	–
Housing status	Non-permanent	0.88	0.56	1.37	0.57
Living assets	High	1	–	–	–
	Average	1.02	0.56	1.87	0.945
	Low	0.76	0.41	1.40	0.384
Working time	Regular time/in shifts	1	–	–	–
	Self-employed	1.74	0.97	3.14	0.066
	≤ 3 million VND	1	–	–	–
Monthly income	> 3 million VND	0.91	0.62	1.33	0.635
	Yes	1	–	–	–
Health care insurance	No	0.29	0.19	0.44	<0.001

Note: Number of observations in the model is 560; p-value of Hosmer-Lemeshow test is 0.36

The regression model in Table 3 was used to estimate the adjusted proportion of health care services among the study populations. Results after adjusting correlates finally show that the utilization of health services in non-migrant populations (67.6%) is better than that in migrant populations. Among migrant populations, seasonal migrants have the lowest proportion of utilization (42%), followed by migrants working in PSE (44%), and then migrants working in IZ (53.7%).

DISCUSSION

The 2009 National Census in Viet Nam categorized internal migrants into four groups: urban-to-urban, urban-to-rural, rural-to-rural, and rural-to-urban migrants depending on their residences and places of origin and destination. The Census also stated rural-to-urban migrants, especially migrants working in industrial zones, account for the majority, compared to other internal migrant populations¹⁹. Additionally, rural-to-urban migrant population also includes a large proportion of migrants working in PSE and seasonal migrants^9,27.

In fact, studies on migration have usually faced the difficulties of sampling representative sample of migrants because of their mobility. However, if a study focused on a certain migrant population alone (e.g. migrants working in industry, seasonal migrants, migrants working on construction sites), sampling would be more feasible, because they usually live in the same areas or in the same collective housing system^28–32. For example, from our study, most seasonal migrants lived in slums, while migrants working in IZ and PSE lived in boarding-houses close to their enterprises. Once a sampling frame is established, it is not too difficult to approach migrants, especially migrants working in IZ and migrants working in PSE because they have regular working schedules (7:30am–5:30pm) or in shifts. We approached migrant groups at their homes with the introduction of a village health worker or a head of a resident unit. Meanwhile, most seasonal migrants are self-employed/freelance and they usually live in slums and have poor security. Therefore, we needed more attempts to approach them with the assistance of a village health worker or a head of a resident unit, but also from the local police. The police only helped researchers reach the place of residence of migrants, and did not participate in the interviews. All these attempts made us approach many more of the migrants, ensure a high response rate, and ensure voluntary participation of migrants.

Many studies identified differences between migrants and non-migrants. Such differences include gender, age, marriage status, living condition, and income^18,19,33. This study not only identified these issues, but also focused on differences among different groups of migrants. For instance, the 2004 Migration Survey and the 2009 National Census stated that most migrants are single and generally younger than non-migrants, but according to this study, this seems not to be consistent with seasonal migrants. Another point is the Survey and Census showed living conditions of migrants to be much worse than non-migrants¹⁸; meanwhile, this study specifically identified seasonal migrants as having the worst living condition, followed by migrants working in PSE and IZ.

The study also showed the health service utilization of non-migrants was better than that of migrants. This is consistent with the 2004 Migration Survey and other studies^8,9,34,35. However, the Survey included all kinds of migrants in all study areas, including Hanoi, the Northeast economic zone, the Central Highlands, Ho Chi Minh City, and the Southeast industrial zone. The percentage of migrants using health care service in this study (about less than 50%) is less than that in the Survey (67.4%) and similar to the proportion in a study by Duong and Liem in 2011^8,9. It should be noted that the definition of “ill health” in this study is the same as in Duong and Liem’s study (i.e. ill enough not to work one day), whereas in the Survey, “ill health” was generally defined as “sick enough to stay home”, but did not specified how many days. Regarding the scale of research, this study involves rural-to-urban migrants in Hanoi, whereas Duong’s and Liem’s study involved rural-to-urban migrants in both Hanoi and Ho Chi Minh City.

Several studies suggested that there are many factors related to the utilization of health care services^36–39. Correlates of the utilization include the supply or availability and accessibility to health care services (i.e. issues of health system), beliefs and attitudes about health care, discrimination and so on. However, among these correlates, some are actually difficult to study and complex to intervene¹³. In Viet Nam, the Migration Survey stated that evidently migration does not bring any pressure to the health system at the destination⁸. Meanwhile, other correlates including gender, age, marriage status, monthly income, education, working hours per day, living standard, and health insurance have also affected health care service utilization¹³. Therefore, this study involved such correlates in order to explore their effects on the utilization of health care services.

Indeed, our study found the utilization of health care services is significantly related to health insurance coverage. Many evidences from the 2004 Migration Survey and other studies in America, Thailand, and China also identified that health insurance is a crucial factor to improve the utilization of health care services^12,13,40,41. As discussed above, previous studies have stated that migration has not brought additional pressures on health system at destination. In other words, the available health system can meet the increase in migration flows and needs of health services for migrants^8,10. Moreover, there are many health facilities (i.e. public and private facilities) at urban settings as this study site and such facilities are available for every person. However, unlike the Migration Survey, the study found minimal or no relation between gender, age, marital status and the utilization⁸. Many previous studies also identified that these correlates, as well as income and education, seem to lead directly to variations in insurance coverage, rather than utilization^9,13. This could be a reason that our study did not find the association of these factors with utilization of health care services.

The correlation between health insurance and the utilization of health care services could partially explain the difference in utilization among migrant populations. This study identified that seasonal migrants use health care services the least, followed by migrants working in PSE and then migrants working in IZ. This is in accordance with the fact that the health insurance coverage of seasonal migrants is also the worst, compared to others. The correlation, once again, illustrates how low health insurance coverage of migrants is. Previous studies found that the insurance coverage of migrants in general was usually less than 50%^8,9,31. Our study specifically showed the coverage is even lower among migrants working in PSE (26%) and seasonal migrants (23%). In contrast, migrants working in IZs are usually provided with health insurance by their employers because they have signed a labor contract with their employers.

Indeed, according to the Viet Nam Insurance Law, employees – both migrants and non-migrants – are all eligible to receive benefits from compulsory health insurance if they have non-term or over 3-month labor contracts. The monthly fee of compulsory health insurance is equal to 4.5% of employees’ monthly salary or allowance, in which employers pay two thirds and employees themselves cover the rest. The salary or allowance used to pay for health insurance is the salary or allowance in the labor contract¹⁵. In other words, employers should automatically extract and pay a health insurance fee from employees’ salary.

In practice, however, employers – especially in small business (e.g. private enterprises, household enterprises, and service business) – tend to avoid paying insurance fee for employees^1,42. Therefore, the Decree No 92/2011/ND-CP was issued to settle administrative violation of regulations on health insurance for employees; but the Decree might not cover seasonal migrants because the migrants mostly work without labor contracts or are self-employed⁴³.

As discussed above, people without compulsory health insurance (e.g. no labor contracts or self-employed) can join voluntary health insurance that should be submitted to and processed by local authority where they have either permanent residence or a temporary residence permit (tam tru – in Vietnamese) and pay the monthly fee themselves¹⁵. Most migrants, especially seasonal migrants, are not willing to buy the health insurance because of their low income and great pressures of supporting their family^9,42. They usually do not register for temporary residence in destination¹. Moreover, even if they have health insurance in their place of origin, they either go back to their hometown’s health center to get a referral for health services, or they directly go to health facilities at the destination but health insurance just pays for a certain proportion of their service fees¹⁵. Such disadvantages are also obstacles for health insurance coverage of migrant populations. Most of migrants working for enterprises in IZs have the labor contracts, thus, they are more likely to have compulsory health insurance compared to other migration groups and even non-migrants.

The study also identified that working time is related to the utilization of health care services. People who have flexible working schedules (i.e. self-employed) are more likely to use health care services than people who work regular hours (i.e. 7:00am – 5:30pm) or in shifts. This might be caused by overlap of working time between enterprises and health service centers. Migrants often do not want to lose a working day to go to health care services because they might lose their salary or bonus. Therefore, they prefer taking medicine at a pharmacy without prescription to seeking health care at health centers^8,42.

Since it is a cross-sectional study, it is difficult to conclude there exists causal relationships between health service utilization and its determinants. However, this is the first study in Viet Nam comparing the utilization of health care services among non-migrant, migrants working in IZ, migrants working in PSE, and seasonal migrants. Results of this study still need sophisticated analysis to better hypothesize causal pathways between determinants and health care utilization. This helps see how these determinants differ among different populations. Moreover, the questionnaire used in this study was validated based on migrants working in industrial zones, but not all migrant groups. Thus, the questionnaire would be not perfectly valid for the remaining groups. Despite these limitations, results of this study can be used for identification of targeted determinants that can be trialed in health interventions.

CONCLUSION

The study compared the utilization of health care services among migrant populations. Non-migrants have the highest proportion of the utilization, whereas migrants have a much lower proportion. Among migrant populations, seasonal migrants have the lowest utilization of health care services, followed by migrants working in industrial zones and migrants working in private small enterprises.

The study also identified health insurance coverage as an important factor related to the utilization of health care services. The coverage of migrants working in IZ is the best while the health insurance coverage is the worst for seasonal migrants. The study pointed out that health insurance coverage needs to be increased if utilization of health care services for the whole population, particularly migrant population is to be improved. Policies of health insurance need to be suitable for migrant population.