Background
With the acceleration of globalization and rapid socioeconomic development in China, domestic migrants have become a powerful driving force of China’s industrialization and urbanization. However, increased population mobility, in particular rural-to-urban migration, has also imposed a great challenge to the provision of healthcare services [
1,
2]. Like in many other countries, there have been increasing concerns in China about the inequality of healthcare services between the migrants and their local counterparts [
3]. Overall, the migrants are less likely to use healthcare services than the non-migrants [
4], in particular in those who are not able to bring their social health insurance entitlements to the migration destinations. This is despite the fact that migrants usually suffer more health problems than non-migrants due to poor living conditions [
5], limited health literacy [
6], overwork [
7], and a lack of social support [
8]. The exposure to a changing environment can add additional risks to the physical and mental health of the migrants [
9].
Massive population migration is also associated with a substantial increase in PM
2.5 emissions [
2]. According to the 2017 China Environmental Bulletin (
https://www.mee.gov.cn/hjzl/zghjzkgb/lnzghjzkgb/), 239 (70.7%) of the 338 municipalities in China failed to meet all of the governmental air quality standards measured by SO
2, NO
2, PM
10, PM
2.5, CO, and O
3. On average, 22% days in a year recorded an air quality index (AQI) below the standards. In China, PM
2.5 is the main air pollutant, which has demonstrated harmful effects on the cardiovascular and respiratory systems [
10,
11]. Compared with windblown soil dusts, PM
2.5 contains more toxic substances like acids and heavy metals, which can penetrate deeper into the lungs of human being [
12]. Long-term exposure to PM
2.5 can increase the incidence of hypertension, respiratory diseases, cardiovascular diseases, and neurological diseases [
13‐
15].
Industrial pollution has been widely perceived as a sizeable health risk by the public [
16]. Children and old people are particularly vulnerable to pollution [
17]. Air pollution is not only associated with physical health harm, but also negative emotion and mental health problems, such as anxiety [
18]. Exposure to air pollution affects the brain function, which can trigger behavioral changes [
19,
20]. There has been increasing awareness of the harm of air pollution among the public. People respond to the problem in a variety of ways. Szyszkowicz reported that air pollution is associated with increased substance abuse [
21]. Chew and colleagues classified individual coping strategies for air pollution into risk aversion, ambiguity aversion, and time decision impatience [
22]. These coping strategies can trigger changes in health behaviors and healthcare seeking behaviors. For example, Yu et al. in a study of school students in China found that air pollution discourages physical activities [
23]. Neidell reported that people living with asthma may reduce outdoor exposure when they perceive heavy air pollution [
24]. Previous studies have been focused on the impact of air pollution on healthcare services relating to respiratory diseases. However, air pollution is also strongly associated with increased cardiovascular morbidity and mortality [
17]. There has been a paucity in the literature documenting the extent of general patients to refrain from seeking healthcare due to concerns of air pollution. Empirical evidence shows that refraining from healthcare can cause serious harm to patients [
25], decreasing quality of life [
4,
26]. It disproportionally impacts people with low socioeconomic status, exacerbating inequity in health and healthcare. Both personal idiosyncrasies and environmental factors play a role in individual decisions of seeking (or refraining from seeking) healthcare services [
27].
This study aimed to address the gap in the literature by determining the incidence of refraining from visiting health facilities in the domestic migrant population in China and its association with air pollution. Internationally, there is widespread inequality in accessibility of healthcare services between migrants and local residents, especially in the low- and middle-income countries [
28], including in China. However, little is known about whether air pollution plays a role in the healthcare-seeking decisions of migrants despite extensive studies into the effects of sociodemographic factors (especially age and income), health insurance entitlements, and seriousness of symptom presentations [
26,
29].
Methods
This study adopted a cross-sectional design using data that are publicly available. No ethics approval was required.
Data source
Health services data were extracted from the 2017 China Migrants Dynamic Survey (CMDS). The CMDS employed a stratified three-stage probability proportionate to size (PPS) sampling strategy to select study participants from the domestic migrant populations in all of the 31 provinces/regions in mainland China. Data were collected through household visits and face-to-face interviews over a one-month period in May 2017. The survey covered data relating to the socioeconomic characteristics (gender, age, education, distance of migration, sense of local belonging, weekly work hour, per capita annual household income) of study participants, their health status (illness/injury/unwell over the past two weeks, self-rated general health, hypertension, diabetes), availability of healthcare resources (social health insurance, distance to nearest health facility), and access to healthcare services (health education, medical services). A total of 169,993 individual responses were returned: only one respondent from each household was allowed. Details of the CMDS methods were published elsewhere (
https://www.chinaldrk.org.cn).
The CMDS data were linked to the air pollution data published by the China National Environmental Monitoring Center (
http://www.cnemc.cn/) according to the residential location of the study participants. The platform publishes real-time weather and air quality information, which include particulate matter (PM
2.5 and PM
10) and air quality index (AQI) at the municipality level in China. Multiple state controlling air sampling sites were set up in each city. Most of the monitoring sites were located in urban areas, with a small number in suburban and rural areas. Details of the environmental conditions around the sampling sites, monitoring methods and analyzers can be available in the previous literature [
30]. The daily average was calculated based on the real-time hourly data captured, while the monthly average was calculated based on the daily average data for the purpose of this study.
The CMDS data were also linked to the atmospheric temperature data of the 310 municipalities in which the study participants resided. The atmospheric temperature data were extracted from the MERRA-2 database released by the National Aeronautics and Space Administration (NASA) in the US. We used the product-M2I6NPANA (doi:
https://doi.org/10.5067/A7S6XP56VZWS), which records the temperature of different layers of air from the surface to 36,000 m high (a total of 42 layers) every six hours. Thermal inversion was defined as a positive temperature difference between the second layer (320 m) and the first layer (110 m) [
31]. We calculated daily average of the thermal inversion strength over the four six-hour periods each day (from 6 pm of the previous day to 6 pm of the current day), and used the daily average to generate monthly average. Each municipality has multiple records of daily average strength of thermal inversion grouped by grids, with each grid covering a 50 × 60 km area. A mean daily value for each municipality was calculated by pooling the records of all relevant grids. A monthly average of the daily mean values was eventually generated for each municipality.
Air quality data processing
We conducted further quality control on the published real-time hourly air quality data (hourly air quality index (AQI), PM
2.5 and PM
10 at the municipality level (
n = 310) in April and May 2017). Referring to previous studies [
32,
33], the original time series were transformed into z scores, and then the unrealistic points in the hourly time series were removed when concurrently satisfying three conditions: (1) an absolute z score larger than 4 (
\(\left|{Z}_{t}\right|>4\)); (2) the increment of the z score between the current time and a previous time larger than 6 (
\({Z}_{t}-{Z}_{t-1}>6\)); and (3) the ratio of the z score to its centred moving average of order 3 larger than 2 (
\({3Z}_{t}/({Z}_{t-1}+{Z}_{t}+{Z}_{t+1})>2\)). The daily average of the above-mentioned indicators were then computed from the hourly data captured when more than 75% of the measurements (18 out of 24) in a day were valid data [
33]. We calculated the monthly average of these indicators for each municipality using the daily average data in April and May 2017 for the purpose of this study.
Measurements
Although the exact dates when the study participants intended to visit health facilities were not known, they fell within the period from 17 April to 31 May 2017. We used three time windows (April average, May average, and April-May average) to test the associations between the air pollution indicators and refraining from visiting health facilities.
We also categorized PM
2.5 and AQI levels into air pollution grades, which align more closely with how people perceive [
37,
38] and adjust their behaviors [
39]. Specifically, the PM
2.5 level was classified as normal (< 35 µg/m
3) or abnormal (≥ 35 µg/m
3) according to the air quality standards (GB3095-2012) in China. These standards have been widely accepted by the public [
40]. AQI represents overall air quality considering all kinds of pollution, which was classified into five grades: good (0–50), moderate (51–100), unhealthy (101–150), very unhealthy (151–200), and hazardous (200+). The five categories were collapsed into three to correct the skewed distribution of data: good (0–50), moderate (51–100), unhealthy (> 100).
Statistical analysis
The proportion of study participants who refrained from visiting health facilities when needed was calculated and compared between those with different characteristics using Chi-square tests.
Probit regression models were established to determine the effect of air pollution on refraining from visiting health facilities after adjustment for variations in the instrumental and control variables (see
Supplementary Table for the coding scheme).
$$Y=a_0+a_1Air\;Pollution+a_j{Control}_j+\varepsilon$$
In the model, Y represents the probability of refraining from visiting health facilities. Air Pollution represents the air quality indicator of the municipality, while Controlj is a set of instrumental and control variables measured at the individual level (j). ε denotes the error term.
Since the intra-cluster (municipality) correlation coefficient (ICC) was low (0.076 < 0.1) indicating a lack of cluster effect [
44], a single-level modeling approach was adopted despite the two levels of measurements. We established the probit regression models using April (Model One), May (Model Two), and April-May PM
2.5 (Model Three), respectively, before introducing the instrumental variable (corresponding monthly thermal inversion intensity) into the models. We also tested the robustness of the findings by replacing PM
2.5 with PM
10 and AQI.
All statistical analyses were performed using Stata/MP 17.0. A two-side p value of lower than 0.05 was considered statistically significant.
Discussion
Although local healthcare institutions can provide better care for the common and frequently-occurring health problems of migrants [
45], we found that more than 16% of the migrant respondents refrained from visiting health facilities when needed over a two-week period. This level of refraining appears to be lower than that (63%) revealed in a survey of domestic migrants in Guangdong province of China in 2012 [
46]. It is also lower than what was found in the general population of Sweden (24%) [
47] and the female undocumented immigrants in the Netherlands (56%) [
48]. It is important to note that many previous studies considered purchase of medicines from community pharmacy retail outlets as part of self-medication, resulting in a much higher percentage of migrants refraining from seeking healthcare [
49]. Self-medication is the top strategy for managing illness conditions in migrant workers in some countries [
50]. In our study, however, pharmacies were considered as one of the health facilities, because the study participants needed to go outside of their houses (being exposed to polluted air) to purchase medicines from pharmacies. Self-medication was reported by almost one quarter (24.32%) of the participants in our study.
Our study offers some new insight into the adverse health effects of air pollution. We found that exposure to ambient air pollutants (PM2.5, AQI, and PM10) is a significant predictor of refraining from visiting health facilities. Other factors that are associated with refraining from visiting healthcare include long-distance (inter-provincial) migration, overwork, good self-rated health, low sense of local belonging, a lack of health education, and absence of hypertension/diabetes.
Negative effect of air pollution on seeking healthcare
We found that air pollution is associated with refraining from visiting health facilities in the domestic migrants in China. A large number of previous studies have revealed negative consequences of air pollution on human health [
51], including in the patients with asthma, chronic obstructive pulmonary disease, and cardiovascular diseases [
17]. Epidemiological and controlled exposure studies show that anyone can suffer from respiratory symptoms and lung function impairment resulting from exposure to air pollution [
52]. People are becoming increasingly concerned about the impacts of air pollution on their health and daily lives [
53]. Air pollution reduces people’s willingness and frequency of travelling [
54]. With easy access to air quality information through the modern information and communication technologies, individuals can easily adjust their health and healthcare-seeking behaviors in line with the level of perceived health risks of air pollution [
55]. A high level of awareness of air pollution and its harm is associated with consumer actions relating to preventive and curative health measures [
21,
56]. These may lead to avoidance of or shortened exposure to external environments [
57]. Empirical evidence shows that the public tend to over-estimate the intensity and hazard of air pollution [
58], causing excessive anxiety and reduced well-being [
40]. The widespread concern about air pollution can also trigger social media sensation [
59], further exacerbating people’s responses. Those who feel threatened by air pollution may refrain from visiting health facilities when needed.
Overwork and refraining from visiting health facilities
People usually migrate to other places in seeking better job opportunities. However, the jobs taken by migrants are often characterized by long working hours, inflexible work schedules, and a lack of job security, which can hinder their access to healthcare services [
60]. Our study provides further evidence to support the association between overwork and refraining from visiting health facilities in the migrant populations. Migrants tend to take heavy workloads, squeezing out of free time for leisure and social activities. In China, there have been increasing concerns about the suboptimal working conditions of domestic migrants, as well as the lower level of social welfare and healthcare entitlements of domestic migrants relative to long-term local residents [
61,
62].
Low social integration and refraining from visiting health facilities
Our study shows that a low sense of local belonging and long-distance (inter-provincial) migration are predictors of refraining from visiting health facilities. Migrants, in particular those who have a temporary or short-term living arrangement, face great challenges in adapting to the new environments and obtaining local social support [
63,
64]. As a result, the continuity of healthcare services needed by migrants can be compromised. Disruption of the existing social networks often makes the migrants feel strange to the health care system of the migrant destinations [
65], jeopardizing their ability to navigate through the often complex health system [
62]. Adding to the complexity are language barriers and socioeconomic discriminations [
66‐
68].
Health education and refraining from visiting health facilities
Health education is key to increasing health literacy, which is usually linked to high use of preventive care [
69]. Our study found that the domestic migrants who received community health education are less likely to refrain from visiting health facilities. This result is consistent with the findings of other studies. Inadequate health literacy is associated with low levels of awareness of healthcare needs and the availability of preventive interventions [
70]. A cross-sectional study of migrant workers in China found that health education is associated with high self-consciousness of seeking healthcare [
4].
Felt health needs play an important role in healthcare seeking decisions in the migrants, according to the findings of our study and some others. We found that those who perceived poor health and suffered from chronic conditions are less likely to refrain from visiting health facilities than others. Shi and colleagues argued that the relatively young and healthy migrants have high capability of ignoring and overcoming health problems [
71]. By contrast, the migrants with moderate and severe symptoms have a more urgent feeling of needs to seek medical care [
72].
Strengths and limitations
This study makes a significant contribution to the literature regarding the health hazards associated with air pollution. The sample size is large. We used three air pollution indicators to determine the association between air pollution and refraining from visiting health facilities when needed. An instrumental variable (strength of thermal inversion) was added to the probit regression modelling to address potential endogeneity.
There are several limitations of this study. First, the dataset used in the study did not tell what illness or how serious the illness conditions were when the study participants refrained from visiting health facilities, although self-rated general health was measured. Second, this study adopted a cross-sectional design, which prevents us from drawing causal conclusions. However, it is unlikely that healthcare seeking behaviors would have a significant impact on air pollution. In China, the vast majority of populations can find a health facility within a 15-minute walking distance. Although we used an instrumental variable to address the endogeneity issue, further studies are recommended to explore the underlying mechanisms of the association between air pollution and refraining from visiting health facilities.
Conclusions and policy implications
This study aimed to determine the association between air pollution and refraining from visiting health facilities in the domestic migrant populations in China. We found that air pollution is a significant predictor of refraining from visiting health facilities. This finding was validated using multiple air pollution indicators, multivariate regression modelling, and inclusion of an instrumental variable.
In addition to air pollution, domestic migrants in China can also be deterred from seeking healthcare due to inequalities in socioeconomic arrangements. Significant regional disparities exist in China, despite the unprecedented economic development over the past few decades [
73]. Regional economic disparities are the key driver of population migration, which is often accompanied by social discrimination against migrants, inequality in welfare entitlements and a low sense of local belonging of the migrants [
63]. Domestic migrants in China have made a great contribution to China’s economic miracle. Governmental policies should pay increasing attention to the wellbeing of the migrant populations, fostering better social integration.
Open AccessThis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit
http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (
http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.
Publisher’s Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.