Setting
China has a longstanding household registration system under which official place of residence is based on place of birth and can only be permanently changed under certain limited circumstances with official permission. Internal migrants thus retain their official resident registration in their sending communities even after moving across administrative boundaries within or between municipal jurisdictions or between provinces. After leaving their homes for a month, they are officially considered migrants or “floaters.” No matter how long ago they moved, most internal migrants are not eligible for many public services, including government-provided health care, in their receiving communities since they don’t have official resident registration at the destination [
27]. An important exception is family planning services, for which all married women of reproductive age are eligible, wherever they live. Health care is also readily available in urban areas from private or public providers, but this must be paid out of pocket without the medical insurance that local urban residents usually have.
Migrant women are eligible for rural health insurance in their sending communities (which requires them to actively enroll and pay nominal premiums). But rural health insurance does not directly cover services outside their home towns. In some circumstances, they can pay out-of-pocket for health care in their receiving communities and submit bills for reimbursement, but the process is cumbersome and reimbursement rates are incomplete and variable. They are not eligible for the urban health insurance that is available to people born in their receiving communities. The Chinese government has for many years had a stated goal of combining the rural and urban health insurance systems, but this has economic and administrative obstacles and there is no specific timeline for implementation.
Migrant women with full time official employment with a labor contract are eligible for separate childbearing insurance included in employee health insurance that covers both maternal medical care and salary for a three to four-month maternity leave and which is mostly paid by the employer, but this does not apply to the majority of female migrant workers who often have part time or temporary jobs or work in the informal sector [
28]. Many migrant women stop working and return to their home towns when they are pregnant, typically from the second trimester through delivery [
29].
Data source
This study examined a subsample of the 2014 National Dynamic Monitoring Survey on Migrants. This large national survey has been conducted every year since 2009 by the China Population and Development Research Center of the National Health and Family Planning Commission (CPDRC). The Investigators applied for and were granted access to the data in 2014. This study could be considered a secondary data analysis in that none of the investigators were directly involved in data collection. The data that we received did not include any personal identifiers. All data were stored in password-protected form.
The Health and Family Planning Commissions of each province undertook the field work for this survey. This included providing basic data for CPDRC to compile the sampling frame, investigator training in the sampled cites, organizing field surveys, data input, data auditing and quality checks. The sampling frame was first initially compiled according to annual reports of internal migrants from each province in 2013 and then refined with new information provided by each province in April 2014. Overall coordination of the survey, including compiling the final sampling frame, sampling, questionnaire design, training of supervisory personnel, field supervision, quality control, data upload, and data management, was directed at the national level.
The total sample was chosen with a stratified, multi-stage and Probability Proportional to Size (PPS) sampling method. Survey information was collected by structured face-to-face interview. Domains included demographic characteristics, employment and family income and expenditure, public health and medical service utilization, childbearing, and family planning service utilization among internal immigrants, both men and women, aged 15–59 throughout China in May 2014. Each of 31 provinces and 1 provincial management unit (XPCC), were regarded as strata. The sample had three stages (sub-district/town, neighborhood or village committee, and individual) and was proportional to the size of the migrant population in each. The sample size by province was also roughly proportional to the size of the migrant population: 14000 interviewees in Zhejiang province, 12,000 in Jiangsu and Guangdong, 10,000 in Heilongjiang, 8000 in Beijing and Shanghai, 7000 in Fujian and Hunan, 6000 in 9 provinces including Tianjin and Shandong, 5000 in 9 provinces including Hebei and Shanxi, and 4000 in Jilin, Guizhou, Tibet, Ningxia, Xinjiang and XPCC. According to the survey plan, 201,000 interviewees were to be interviewed; the final sample size was 200,937.
Supervisors in charge of quality control and training at the provincial level were trained at the national level. Uniform survey manuals and investigator handbooks were printed and distributed to each province by CPDRC. In the field, investigators visited the sampled migrants and interviewed them for the survey. Each interview took about 20–30 min. Potential participants received a printed information sheet explaining the study and informing them that participation is voluntary. They then either declined or gave verbal consent to participate. Refusals were rare (as is common for such surveys in China) and were not tabulated; sampled individuals who did not participate were replaced. Completed questionnaires were checked for quality by supervisors. Each province scheduled and finished the survey in May 2014. Further details of the overall survey methodology are available in the “2015 Report on China’s Migrant Population Development.” [
30]
This study was conducted on a subset of the survey described above. Criteria included a rural Hukou (household registration), female, married, aged 20–34, and having delivered a baby since January of 2012. This subset thus included women giving birth at ages as young as 17 years 8 months and accounted for 93.6% of all women in the larger survey from rural Hukous who gave birth in the previous two years. The resulting sample size was 5372. Prenatal care utilization data were considered for the youngest child if a respondent had more than one delivery since 2012.
Independent variables
Two main groups of independent variables were analyzed: those related to demographic characteristics and those related to migration characteristics. For the demographic characteristics, age was divided into 3 categories: 20–24 years, 25–29 years, and 30–34 years; educational level was asked with 7 options (illiterate, elementary, middle school, high school, junior college, undergraduate, graduate) and was recoded into 4 categories: elementary school, middle school, high school, or college; area of origin was recoded into 3 categories: east, central, and west (these are recognized regions of China with substantial differences in development); birth order of this child was coded 1st, 2nd, and 3rd or more (only 17 migrant women had 4 children and one woman had 5); having medical insurance was recoded into 4 categories: rural medical insurance, childbearing insurance, both insurances, and no insurance; household monthly income per capita (CNY) was divided into 4 categories: < 1000, 1000–2000, 2000–3000, and > =3000. For migration characteristics, migration before this pregnancy was coded yes or no (some women did not first migrate until after their recent pregnancy); geographic spread of migration was recoded into 3 categories: between provinces (from a province of origin to another province), between municipal jurisdictions within province (from a city of origin to another city in a province), and within municipal jurisdiction (municipal jurisdictions in China often include large surrounding rural areas.) Staying at one’s rural hometown during pregnancy was coded yes or no. Time since first migration (in years) was a numerical variable calculated with 2 time variables: time of first migration and time of the survey.