Changes in caesarean section rates in China during the period of transition from the one-child to two-child policy era: cross-sectional National Household Health Services Surveys.

OBJECTIVES: Since 2009, China has introduced policies, principally targeting health professionals, to reduce caesarean section (CS) overuse. In 2016, China endorsed a universal two-child policy. Advanced maternal age and previous CS may indicate changes in obstetric risks, which raise concerns on the need for and safety of CS. This study investigated changes in CS rates in 2008-2018, and factors associated with CS use during the period of transition from the one-child to two-child policy era.
DESIGN: We used births data from the cross-sectional National Household Health Services Surveys in 2013 and 2018.
SETTING: Population-based national survey. PARTICIPANTS: Women who had the last live birth within 5 years before the survey. PRIMARY OUTCOME MEASURE: CS rate.
RESULTS: Overall CS use increased from 40.9% in 2008 to 47.2% in 2014 with significant increase in rural areas and the western region, and slightly decreased to 45.2% in 2018 with the greatest decrease among nulliparous women. Maternal request for CS by urban nulliparous women decreased from 36.8% in 2008-2009 to 22.2% in 2016-2018, but this change was not statistically significant in rural areas. Maternal age over 35 years old (OR 2.40, 95% CI 1.72 to 3.35) and births that occurred at a private hospital (OR 1.52, 95% CI 1.25 to 1.86) were associated with CS use among nulliparous women in 2016-2018. The CS rate among multiparous women increased over time. Individual socioeconomic factors associated with CS use among multiparous women.
CONCLUSIONS: The CS rate rise in China in 2008-2018 is attributable to increased use in rural areas and the less developed western region. The population policy shift, alongside facility policies for unnecessary CS reduction, are likely factors in CS reduction in urban areas. The challenge remains to reduce unnecessary CS, at the same time as providing safe, universal access to CS for women in need. © Author(s) (or their employer(s)) 2022. Re-use permitted under CC BY-NC. No commercial re-use. See rights and permissions. Published by BMJ.

This study used well-established population-based national survey data to examine the change of caesarean section (CS) rate in China by urban and rural areas, across regions and women’s characteristics over the periods of population policy shift.

This study could not separate the effects of strategies to reduce unnecessary CS and the shift of the population policy on the use of CS in China.

We are not able to conduct more subgroup analysis such as previous CS or CS with or without medical indications due to unavailable data.

Introduction

Globally, caesarean section (CS) rates are rising in all regions with one-fifth of live births being by CS in 2015.1 Complex social, cultural, economic and medical factors are known drivers of CS use.2 Overuse and underuse of CS represent simultaneous challenges for many health systems. Overuse of CS, where CS is performed without or on the basis of ambiguous medical indications, has been associated with increased risk of maternal and newborn adverse outcomes and increased costs for health systems and individuals.3 4 Meanwhile, the low use of CS implies limited accessibility to this life-saving procedure for women in need during childbirth.1 The WHO Statement on Caesarean Section Rates suggests that CS rates at population level higher than 10% are not associated with reductions in maternal and newborn mortality rates, while every effort should be made to provide CS to women in need.5

China has witnessed a rapid increase in the use of CS since 1990s.6–9 According to the data from the National Household Health Services Surveys in China, the CS rate increased from 19.2% in 2003 to 36.3% in 2011.7 Many CSs in China are not medically indicated.10 There is evidence that individual, health system and sociocultural factors are driving the use of CS.11 Women may request CS because of fear of labour pain, fear of risk and adverse outcomes of vaginal delivery, perceived convenience of CS for birth plan, and control and perceived CS as a safe option for childbirth.11 In China, facility-based delivery is a national strategy to reduce risks of adverse outcomes for mothers and newborns. Almost all births occurred in health facilities by 2015. It has been argued that financial incentive and fear of malpractice may shape the preference of health professionals for performing CS in the hospital settings.11 12

Since 2009, the Chinese government has increasingly introduced policies and strategies at national and local levels to restrict the use of unnecessary CS.9 These strategies largely targeted health professionals. They include revising clinical guideline to strict control of CS indications, strengthening training of midwifery care and audit of CS without medical indications, setting facility CS rate targets and removing financial incentives for CS. China has gradually relaxed their family planning policy since 2013, with all families being allowed and encouraged to have a second child in 2016. Advanced maternal age and previous CS may indicate changes in obstetric risks, which raise concerns on the need for and safety of CS.13 14

Recent studies that used data from the National Maternal Near Miss Surveillance System (NMNMSS) reported a moderate decrease of CS rate in some big cities, which coincided with the period of relaxation of the one-child policy between 2012 and 2016.8 9 This decrease in CS rate may be attributable to facility strategies to reduce the use of CS without medical indications. However, the authors also acknowledged the limitation of NMNMSS data in over-representing urban populations. Little is known about trends of CS use in rural areas, across regions at different stages of socioeconomic development, or how facility strategies to mitigate unnecessary CS and the population policy shift have affected the use of CS in these places.

This study used cross-sectional data from the National Household Health Services Surveys conducted in 2013 and 2018, which achieved reliable representativeness of the general population by urban and rural areas and across socioeconomic development regions. We investigated changes in CS rates between 2008 and 2018, by urban and rural location, and across socioeconomic regions in China. It sought to examine maternal request for CS by the study periods and by parity, as well as demographic and socioeconomic factors associated with use of CS during the period of transition from the one-child to two-child policy era.

Methods

Data source

We obtained the permission to access the birth dataset from the National Household Health Services Surveys conducted in 2013 and 2018. Each survey employed the same three-stage, stratified, cluster random sampling procedure. At the first stage, urban and rural location and socioeconomic regions were used to classify cities and counties into six groups: eastern urban, eastern rural, central urban, central rural, western urban and western rural. Simple random sampling was used to select cities and counties from each group. The random sample process was repeated for three times to select the ones most close to the parameters (eg, fertility rate, mortality rate, demographic structure, etc) representing the general. Then five subdistricts or townships were randomly selected from each city or county based on the rank of the number of population. Finally, three communities or villages from each subdistrict or township were randomly selected and all households in the selected subdistrict or township were included in the survey. In total, 93 613 households were included in the survey of 2013 and 94 074 in the survey of 2018.

The trained primary health workers administered face-to-face survey to each family member in the sampled households using structured questionnaire. The questionnaires used in the two surveys had the same structure and involved similar questions, which included the general demographic and socioeconomic characteristics of the sampled households and family members, and the utilisation of and expenditures on health services. There is one section on the childbirth that asked questions about the use of antenatal care, place of delivery, mode of delivery and caesarean delivery for maternal request. We included women who had the last live birth within 5 years before the survey in this study to avoid over-representation of women who have one more child.

Measures

The outcome measure was CS rate, the percentage of births by CS. In the survey, the mode of delivery was asked with the following question: ‘How did the birth take place: (a) vaginal delivery; (b) CS’. If the answer was ‘CS’, the following question was ‘Who was the most important person to propose CS: (a) myself; (b) husband; (c) parent; (d) doctor; (e) others’. We considered CS as a woman request in the analysis if the woman chose the option ‘a (myself)’.

We examined demographic and socioeconomic factors associated with the use of CS that included: maternal age (<25, 25–34, ≥35 years); maternal educational level (illiterate or primary school, secondary school, high school/professional school or higher); location of residence (urban, rural); living in different socioeconomic region (developed eastern, less developed central, least developed western); health insurance coverage; income quartile; parity, defined as the number of live births born by a woman; and place of delivery, defined as type of health facility where the live birth occurred (county or higher-level hospital, maternal and child health hospital, township/community health centre, private hospital). There are three basic health insurance schemes in China: Urban Employee Basic Health Insurance (UEBMI), Urban Residents Basic Health Insurance (URBMI) and rural New Cooperative Medical Scheme (NCMS). In recent years, URBMI was integrated with NCMS in some provinces renamed as Urban and Rural Residents Basic Medical Insurance (URRBMI). Overall, UEBMI provides better coverage for both inpatient and outpatient care compared with URRBMI. In the analysis, we grouped URBMI, NCMS and URRBMI into one category as ‘URRBMI’. Health insurance coverage was grouped into: none coverage, UEBMI, URRBMI and others (including free medical service scheme for special sectors or labour insurance). Annual household income in the calendar year that preceded the survey included savings and household expenditure on consumables during that year. We generated income quartile by dividing household income by the number of individuals in the household, which reflected the lowest-income group (quartile 1), low-income group (quartile 2), middle-income group (quartile 3) and high-income group (quartile 4).

Data analysis

We investigated changes of CS rate in 2008–2018 by urban and rural areas and across different socioeconomic regions. We also examined CS rate among nulliparous and multiparous women by location and region in the study periods. We studied change of women request for CS by parity that the time period 2008–2018 was divided into 2008–2009, 2010–2012, 2013–2015 and 2016–2018. Χ2 test was used to test the difference by the study period. We conducted bivariate and multivariate logistic regression analyses to study explanatory variables associated with the use of CS for all and in stratification of urban and rural areas and by socioeconomic regions in 2008–2018. In addition, we stratified data in 2016–2018 to study demographic and socioeconomic factors associated with the use of CS after the universal two-child policy in China for all and by nulliparous and multiparous women, by location and regions. We applied Stata V.13.0 for data analysis.

Patient and public involvement

No patient involved.

Results

A total of 23 053 women who had a live birth in the study period 2008–2018 were included in the analysis (table 1). The distribution of maternal age was relatively similar by the study periods, and more than 60% of women were aged 25–34 years. The proportion of women who attended high school and professional school or higher and those who had two or more children increased over time. Few women had no health insurance coverage, and a vast majority of women were enrolled in URRBMI. In addition, there were more women in urban areas giving birth than women in rural areas observed in the period of 2016–2018. The distribution of women living in a region was relatively similar over time. The majority of women gave birth in a general hospital (county or higher-level hospital), and this proportion increased over time. The proportion of women giving birth in community and township health centres decreased in both urban and rural areas across different socioeconomic regions (online supplemental table 1). Few women reported giving birth in a private hospital in the survey of 2018, which was not reported in the survey of 2013.

Table 1

Demographic and socioeconomic characteristics of women giving birth in China, 2008–2018

Characteristics	2008–2009(n=2638)	2010–2012(n=7015)	2013–2015(n=6151)	2016–2018(n=7249)	Total(n=23 053)
	No (%)	No (%)	No (%)	No (%)	No (%)
Age
<25	146 (5.5)	1395 (19.9)	785 (12.8)	961 (13.3)	3287 (14.3)
25–34	1708 (64.7)	4514 (64.3)	3836 (62.4)	4775 (65.9)	14 833 (64.3)
≥35	784 (29.7)	1106 (15.8)	1530 (24.9)	1513 (20.9)	4933 (21.4)
Educational level*
Illiterate or primary school	493 (18.7)	1091 (15.6)	869 (14.1)	671 (9.3)	3124 (13.5)
Secondary school	128 (48.7)	337 (48.0)	250 (40.8)	250 (34.5)	9662 (41.9)
High school/professional school or higher	861 (32.6)	2554 (36.4)	2773 (45.1)	4078 (56.3)	10 266 (44.5)
Parity†
1	1424 (54.0)	4068 (58.0)	2965 (48.2)	2937 (40.5)	11 394 (49.4)
≥2	1213 (46.0)	2947 (42.0)	3184 (51.8)	4312 (59.5)	11 656 (50.6)
Health insurance coverage
None	115 (4.4)	366 (5.2)	334 (5.4)	386 (5.3)	1201 (5.2)
UEBMI	470 (17.8)	1112 (15.9)	1343 (21.8)	1953 (26.9)	4878 (21.2)
URRBMI	2007 (76.1)	5440 (77.5)	4270 (69.4)	4589 (63.3)	16 306 (70.7)
Others	46 (1.7)	97 (1.4)	204 (3.3)	321 (4.4)	668 (2.9)
Location
Urban	1234 (46.8)	3261 (46.5)	3133 (50.9)	4166 (57.5)	11 794 (51.2)
Rural	1404 (53.2)	3754 (53.5)	3018 (49.1)	3083 (42.5)	11 259 (48.8)
Region
Eastern	878 (33.3)	2238 (31.9)	2133 (34.7)	2741 (37.8)	7990 (34.7)
Central	875 (33.2)	2309 (32.9)	1822 (29.6)	1957 (27.0)	6963 (30.2)
Western	885 (33.5)	2468 (35.2)	2196 (35.7)	2551 (35.2)	8100 (35.1)
Place of delivery‡
County or higher-level hospital	1416 (53.7)	4163 (59.3)	3755 (61.0)	4746 (65.5)	14 080 (61.1)
Maternal and child health hospital	670 (25.4)	1743 (24.8)	1482 (24.1)	1678 (23.1)	5573 (24.2)
Township/community health centre	552 (20.9)	1109 (15.8)	633 (10.3)	391 (5.4)	2685 (11.6)
Private hospital	—	—	281 (4.6)	434 (6.0)	715 (3.1)

Others include free medical service scheme for special sectors or labour insurance.

*Data were missing for one woman in 2013–2015.

†Data were missing for one woman in 2008–2009, and two in 2013–2015.

‡The private hospital was not included in the survey in 2013.

UEBMI, Urban Employee Basic Medical Insurance; URRBMI, Urban and Rural Resident Basic Medical Insurance.

CS rate

Nationwide, the overall CS rate increased from 40.9% in 2008 to 47.2% in 2014. After the scale-up of the two-child policy, the CS rate slightly decreased; it was 45.2% in 2018.

CS rate by urban and rural areas and across regions

In urban areas, the CS rate slightly increased from 50.4% in 2008 to 52.3% in 2014, and then slightly decreased to 47.8% in 2018. However, in rural areas, the CS rate had significantly increased from 33.1% in 2008 to 43.8% in 2015. In rural areas, there was also a slight decrease after the relaxation of the one-child policy. The CS rate in rural areas was 41.2% in 2018 (figure 1A). A similar trend was found across different socioeconomic regions (figure 1B). We observed a significant increase in CS rate from 28.1% in 2008 to 38.6% in 2018 in the least developed western region. In the stratification of urban and rural areas by regions, in 2008 the CS rates in urban areas in the eastern, central and western regions were 50.9%, 62.3% and 37.4%, respectively. The difference of CS rate in urban areas by region became very small in 2018 (48.1% in eastern, 49.2% in central and 46.6% in western region). The CS rates in rural areas across all regions increased between 2008 and 2018. The CS rates in the eastern and central rural areas were higher or close to the rate in urban areas in these two regions (online supplemental figure 1).

Figure 1

Proportion of women giving birth by caesarean section by urban and rural (A) and across regions (B), 2008–2018. Caesarean section rate in urban areas slightly decreased between 2008 and 2018, while it had increased in rural areas. Across regions, a large increase occurred in the less developed western region.

CS rate by parity

Around half of nulliparous women in urban areas gave birth by CS between 2008 and 2014, and the proportion in rural areas grew significantly from 33.3% in 2008 to 49.6% in 2015. The CS rate among nulliparous women decreased rapidly in both urban (42.8% in 2018) and rural areas (37.5% in 2018) after the universal two-child policy (figure 2A). The CS rate among multiparous women continued to increase from 35.3% in 2008 to 48.4% in 2018 with similar trends in both urban and rural areas (figure 2B). We found the similar finding in terms of the change of CS rate by parity across different socioeconomic regions (online supplemental figure 2).

Figure 2

Proportion of women giving birth by caesarean section by nulliparous (A) and multiparous women (B) by urban and rural areas, 2008-2018. Caesarean section rate among nulliparous women decreased around 2016 in both urban and rural areas, while there was a large increase in the number of multiparous women delivering by caesarean section.

Maternal request for CS

We examined maternal request for CS by the study periods (table 2). According to women’s self-report, the proportion of maternal request for CS among nulliparous women decreased from 35.8% in 2008–2009 to 24.4% in 2016–2018 (p<0.01). In the stratification of residents’ location, maternal request for CS significantly decreased in urban areas from 36.8% in 2008–2009 to 22.2% in 2016–2018 (p<0.01); however, the change in rural areas was not statistically significant (from 33.5% in 2008–2009 to 29.4% in 2016–2018) (online supplemental table 2). Among multiparous women, around one-third of women reported maternal request for CS, and there was no significant change between 2008–2009 and 2016–2018.

Table 2

Proportion of women who had caesarean section (CS) reporting recommendation by others and own request for CS by parity in China, 2008–2018 (%)

Characteristics	2008–2009	2010–2012	2013–2015	2016–2018	P value
	n=1059	n=2912	n=2794	n=3208
Parity 1
Women request	35.8	30.2	27.9	24.4	<0.001
Husband	0	0	1.3	1.6	<0.001
Doctor	63.3	68.8	69.5	72.7	<0.001
Others	0.9	1.0	1.3	1.3	0.877
Parity ≥2
Women request	31.7	30.0	32.6	30.9	0.445
Husband	0	0	1.8	1.2	<0.001
Doctor	66.3	69.1	64.4	67.1	0.094
Others	2.0	0.9	1.2	0.8	0.243

In addition, the proportion of CS suggested by a doctor among nulliparous women increased from 63.3% in 2008–2009 to 72.7% in 2016–2018 (p<0.01), and there was no significant change in doctors’ suggestion for CS among multiparous women by the study period. For both nulliparous and multiparous women, there were few CSs proposed by women’s husband and others (table 2).

Demographic and socioeconomic factors associated with the use of CS

In the study period of 2008–2018

Table 3 shows factors associated with the use of CS in China by urban and rural areas over the study period of 2008–2018. After adjusting for all explanatory variables, the use of CS was less common in urban areas in the survey of 2018 (OR 0.85, 95% CI 0.78 to 0.92) compared with the survey of 2013; however, it was more common in rural areas (OR 1.30, 95% CI 1.19 to 1.41) in the survey of 2018. Advanced maternal age (≥35 years), having secondary education or higher, and giving birth at high-level hospital or private hospital were significantly associated with the use of CS in both urban and rural areas. In rural areas, women from the highest-income quartile were more likely to have CS (OR 1.69, 95% CI 1.47 to 1.95) compared with women from the lowest quartile, and multiparous women were less likely to have CS (OR 0.80, 95% CI 0.73 to 0.88) than nulliparous women. However, these differences were not statistically significant in urban areas.

Table 3

Factors associated with use of caesarean section in China by location, 2008–2018

	All		Urban		Rural
	Unadjusted	Adjusted*	Unadjusted	Adjusted*	Unadjusted	Adjusted*
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Year of the survey
2013	1.00	1.00	1.00	1.00	1.00	1.00
2018	1.16(1.10 to 1.22)	1.03(0.97 to 1.09)	0.92(0.86 to 0.99)	0.85(0.78 to 0.92)	1.38(1.27 to 1.49)	1.30(1.19 to 1.41)
Age
<25	1.00	1.00	1.00	1.00	1.00	1.00
25–34	1.44(1.33 to 1.55)	1.32(1.21 to 1.43)	1.29(1.14 to 1.46)	1.29(1.13 to 1.47)	1.36(1.23 to 1.51)	1.35(1.20 to 1.52)
≥35	2.03(1.85 to 2.22)	2.02(1.82 to 2.25)	1.93(1.68 to 2.21)	2.05(1.75 to 2.39)	1.78(1.57 to 2.02)	2.00(1.72 to 2.32)
Educational level
Illiterate or primary school	1.00	1.00	1.00	1.00	1.00	1.00
Secondary school	1.32(1.21 to 1.43)	1.21(1.10 to 1.32)	1.10(0.94 to 1.29)	1.16(0.98 to 1.36)	1.32(1.19 to 1.46)	1.14(1.02 to 1.27)
High school or higher	1.76(1.62 to 1.92)	1.21(1.10 to 1.34)	1.21(1.05 to 1.41)	1.22(1.03 to 1.44)	1.66(1.48 to 1.86)	1.10(0.96 to 1.26)
Residence
Urban			—	—	—	—
Rural	0.61(0.58 to 0.65)	0.75(0.70 to 0.80)	—	—	—	—
Region
Eastern	1.00	1.00	1.00	1.00	1.00	1.00
Central	1.18(1.11 to 1.26)	1.25(1.17 to 1.34)	1.38(1.26 to 1.51)	1.35(1.23 to 1.48)	1.01(0.92 to 1.11)	1.12(1.01 to 1.23)
Western	0.62(0.58 to 0.66)	0.69(0.64 to 0.73)	0.91(0.83 to 0.99)	0.92(0.84 to 1.01)	0.43(0.39 to 0.47)	0.48(0.43 to 0.53)
Health insurance coverage
URRBMI	1.00	1.00	1.00	1.00	1.00	1.00
UEBMI	1.52(1.42 to 1.62)	0.98(0.91 to 1.07)	1.11(1.03 to 1.20)	1.01(0.92 to 1.11)	1.91(1.63 to 2.24)	1.13(0.94 to 1.35)
None	1.25(1.11 to 1.40)	0.96(0.85 to 1.09)	0.94(0.82 to 1.08)	0.91(0.79 to 1.05)	1.42(1.13 to 1.79)	1.27(1.00 to 1.61)
Others	1.34(1.15 to 1.57)	1.05(0.90 to 1.24)	1.18(0.97 to 1.43)	1.12(0.92 to 1.36)	1.12(0.84 to 1.47)	1.06(0.79 to 1.42)
Income quartile
Quartile 1	1.00	1.00	1.00	1.00	1.00	1.00
Quartile 2	1.39(1.29 to 1.50)	1.24(1.15 to 1.34)	1.43(1.30 to 1.58)	1.10(0.97 to 1.24)	1.43(1.30 to 1.58)	1.28(1.15 to 1.41)
Quartile 3	1.62(1.50 to 1.74)	1.30(1.21 to 1.41)	1.63(1.47 to 1.80)	1.14(1.02 to 1.29)	1.63(1.47 to 1.80)	1.36(1.22 to 1.52)
Quartile 4	1.76(1.63 to 1.89)	1.26(1.16 to 1.38)	2.04(1.79 to 2.32)	1.07(0.95 to 1.21)	2.04(1.79 to 2.32)	1.69(1.47 to 1.95)
Parity
1	1.00	1.00	1.00	1.00	1.00	1.00
≥2	0.87(0.83 to 0.92)	0.87(0.82 to 0.93)	1.02(0.95 to 1.10)	0.94(0.86 to 1.02)	0.89(0.83 to 0.96)	0.80(0.73 to 0.88)
Place of delivery
Township/community health centre	1.00	1.00	1.00	1.00	1.00	1.00
County or higher-level hospital	2.23(2.03 to 2.44)	2.03(1.85 to 2.24)	1.57(1.36 to 1.81)	1.48(1.27 to 1.72)	2.49(2.21 to 2.81)	2.45(2.16 to 2.78)
Maternal and child health hospital	2.02(1.83 to 2.23)	1.76(1.59 to 1.96)	1.41(1.21 to 1.64)	1.34(1.14 to 1.57)	2.02(1.75 to 2.32)	1.98(1.71 to 2.29)
Private hospital	2.55(2.15 to 3.02)	2.31(1.93 to 2.76)	1.94(1.53 to 2.47)	1.91(1.49 to 2.46)	2.46(1.92 to 3.17)	2.44(1.87 to 3.18)

Others include free medical service scheme for special sectors or labour insurance.

*Adjusting for all explanatory variables.

UEBMI, Urban Employee Basic Medical Insurance; URRBMI, Urban and Rural Residents Basic Medical Insurance.

Across different socioeconomic regions, the use of CS increased in the western region in the survey of 2018 (OR 1.10, 95% CI 1.00 to 1.22) compared with the survey of 2013, while this difference was not statistically significant in eastern and central regions (online supplemental table 3). Advanced maternal age (≥35 years) and giving birth at a high-level hospital or private hospital were associated with the use of CS in all regions. In central and western regions, women who lived in rural areas, were from low-income quartile household and had more than one child were less likely to have CS.

In the study period of 2016–2018

Following the universal two-child policy in China (2016–2018), women in urban areas who were at advanced age (≥35 years), multiparous and gave birth at a county or high-level hospital or private hospital were more likely to have CS (online supplemental table 4). In rural areas, in addition to maternal age and place of delivery, maternal education attainment and household income were also positively associated with the use of CS. Factors associated with the use of CS in eastern and central regions at the same study period were found similar in urban areas and those in the western region were found similar in rural areas (online supplemental table 5).

We examined the factors associated with the use of CS by parity in the era of the two-child policy. We found that advanced maternal age (≥35 years) and births that occurred at a private hospital were significantly associated with the use of CS among nulliparous women after adjusting for all explanatory variables, while the association between the use of CS and other socioeconomic factors was not statistically significant (table 4). For multiparous women, women who were older, had higher educational attainment, had health insurance coverage, were from a wealthy household, lived in urban areas or the eastern region, and gave birth at a higher-level hospital or private hospital were more likely to have CS after adjusting for all explanatory variables (table 4).

Table 4

Factors associated with use of caesarean section after relaxation of the one-child policy in China by parity, 2016–2018

	All		Parity 1		Parity ≥2
	Unadjusted	Adjusted*	Unadjusted	Adjusted*	Unadjusted	Adjusted*
	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
Age
<25	1.00	1.00	1.00	1.00	1.00	1.00
25–34	1.78(1.53 to 2.06)	1.62(1.38 to 1.90)	1.52(1.27 to 1.83)	1.52(1.25 to 1.86)	2.37(1.79 to 3.14)	2.06(1.54 to 2.75)
≥35	2.89(2.44 to 3.43)	2.58(2.12 to 3.13)	2.41(1.76 to 3.31)	2.40(1.72 to 3.35)	3.73(2.78 to 5.00)	3.19(2.34 to 4.33)
Educational level
Illiterate or primary school	1.00	1.00	1.00	1.00	1.00	1.00
Secondary school	1.40(1.18 to 1.67)	1.42(1.18 to 1.71)	0.96(0.65 to 1.41)	1.02(0.69 to 1.51)	1.58(1.30 to 1.93)	1.49(1.21 to 1.84)
High school or higher	1.38(1.17 to 1.64)	1.24(1.02 to 1.51)	0.97(0.67 to 1.39)	0.85(0.58 to 1.25)	1.74(1.43 to 2.12)	1.35(1.08 to 1.70)
Residence
Urban	1.00	1.00	1.00	1.00	1.00	1.00
Rural	0.78(0.71 to 0.86)	0.84(0.75 to 0.95)	0.81(0.69 to 0.95)	0.85(0.70 to 1.02)	0.72(0.63 to 0.81)	0.84(0.73 to 0.97)
Region
Eastern	1.00	1.00	1.00	1.00	1.00	1.00
Central	1.01(0.90 to 1.13)	1.01(0.90 to 1.14)	0.99(0.82 to 1.19)	1.03(0.84 to 1.25)	1.01(0.87 to 1.17)	1.00(0.86,1.17)
Western	0.65(0.58 to 0.73)	0.70(0.62 to 0.78)	0.83(0.69 to 0.98)	0.87(0.72 to 1.04)	0.55(0.48 to 0.64)	0.61(0.52 to 0.71)
Health insurance coverage
URRBMI	1.00	1.00	1.00	1.00	1.00	1.00
UEBMI	1.19(1.07 to 1.32)	0.95(0.83 to 1.09)	1.17(1.00 to 1.37)	1.05(0.86 to 1.28)	1.34(1.15 to 1.55)	0.91(0.76 to 1.10)
None	0.92(0.75 to 1.14)	0.81(0.64 to 1.01)	1.07(0.77 to 1.49)	0.99(0.70 to 1.38)	0.85(0.65 to 1.13)	0.69(0.51 to 0.93)
Others	1.14(0.91 to 1.44)	0.97(0.76 to 1.23)	0.81(0.54 to 1.22)	0.76(0.50 to 1.14)	1.36(1.03 to 1.81)	1.08(0.80 to 1.47)
Income quartile
Quartile 1	1.00	1.00	1.00	1.00	1.00	1.00
Quartile 2	1.22(1.07 to 1.39)	1.12(0.97 to 1.28)	1.12(0.89 to 1.41)	1.07(0.84 to 1.35)	1.28(1.09 to 1.51)	1.12(0.95 to 1.33)
Quartile 3	1.44(1.27 to 1.64)	1.24(1.07 to 1.43)	1.10(0.89 to 1.37)	0.99(0.79 to 1.25)	1.78(1.51 to 2.10)	1.40(1.17 to 1.68)
Quartile 4	1.28(1.13 to 1.46)	1.04(0.89 to 1.21)	1.13(0.92 to 1.39)	0.94(0.73 to 1.19)	1.51(1.27 to 1.79)	1.10(0.90 to 1.35)
Parity
1	1.00	1.00	—	—	—	—
≥2	1.29(1.18 to 1.42)	1.12(1.01 to 1.25)	—	—	—	—
Place of delivery
Township/community health centre	1.00	1.00	1.00	1.00	1.00	1.00
County or higher-level hospital	2.03(1.62 to 2.54)	2.09(1.66 to 2.64)	1.36(0.85 to 2.19)	1.27(0.78 to 2.06)	2.45(1.89 to 3.16)	2.34(1.80 to 3.05)
Maternal and child health hospital	1.67(1.32 to 2.12)	1.68(1.32 to 2.15)	1.15(0.70 to 1.88)	1.03(0.62 to 1.70)	1.99(1.51 to 2.62)	1.88(1.42 to 2.51)
Private hospital	2.16(1.62 to 2.88)	2.26(1.68 to 3.04)	1.45(0.83 to 2.54)	1.52(1.25 to 1.86)	2.56(1.81 to 3.61)	2.55(1.79 to 3.64)

Others include free medical service scheme for special sectors or labour insurance.

*Adjusting for all explanatory variables.

UEBMI, Urban Employee Basic Medical Insurance; URRBMI, Urban and Rural Residents Basic Medical Insurance.

Discussion

Summary of key findings

In China, the CS rate increased between 2008 and 2015, which was, to a great extent, attributable to a rapid increase of the use of CS in rural areas and the least developed western region. After the scale-up of the two-child policy, the CS rate slightly decreased in both urban and rural areas and across socioeconomic regions, particularly among nulliparous women. The proportion of maternal request for CS decreased among nulliparous women in urban areas over time; however, this proportion decreased slightly in rural areas that 30% of women underwent CS due to maternal request for CS in 2016–2018. In the era of the two-child policy, advanced maternal age and births that occurred in a private hospital were associated with the use of CS among nulliparous women. The CS rate among multiparous women continued to increase over time, and demographic and socioeconomic factors were positively associated with the use of CS among multiparous women.

Strengths and limitations

This study contributes to what is known about rates of CS in China, where most existing studies are limited to a few hospitals or regions. It is a strength of this paper that with the increase of population size and urbanisation in China over the past two decades, the National Household Health Services Surveys adapted their sampling method in 2013 and increased the sample size to achieve reliable representativeness of the general population by urban and rural areas and across different socioeconomic regions. It provides unique insights into both mode of birth and whom households report proposed actual caesarean births. However, several limitations in terms of data and analysis remain. First, all information was based on women’s reports, and the reasons for maternal request or doctor suggestion for CS were not asked. We were not able to distinguish in this study how many CSs performed were medically indicated. Second, women’s history of pregnancy (eg, previous CS or others) was not available. We could not make a subgroup analysis on the use of CS among women with or without uterine scar. Third, we could not separate the effects of strategies to reduce unnecessary CS and the shift of the population policy on the use of CS in China. That said, we did observe a slight decrease of CS rate in line with the period of the universal two-child policy, and the CS rate in urban areas and regions with a high baseline rate had a moderate change over time (2008–2018). Thus, the interpretation of a decrease in CS rate following relaxation of the one-child policy should be made with caution.

Interpretations

We observed a rapid decrease of CS rate among nulliparous women in both urban and rural areas and across all regions in line with the period of the universal two-child policy, which is consistent with the findings in other studies.9 15 However, any causal association remains speculative, not least because of the nuances within the one-child policy itself. In 1979, China announced its family planning policy to strictly control population size. The policy included rules of governing marriage, contraception, number of births and spacing where a second child was permitted.16 The one-child rule was strictest for urban residents and employees of the government agencies. In rural areas, a second child was generally allowed after 5 years, especially if the first born was a girl. Some ethnic minorities were permitted a third child. With socioeconomic development and change of demographic structure, the Chinese government gradually relaxed the one-child policy over a decade with the entire population encouraged to have a second child since 2016.17

The CS rate decreased moderately in urban areas and the eastern and central regions, which had a relatively high baseline CS rate. One plausible explanation is that this may be attributable to the introduction of policies and strategies aiming to reverse the high CS rate through a national top-down approach in China, although results of introducing comprehensive interventions to mitigate unnecessary CSs are mixed in previous studies.8 9 18

We found that the CS rate increased dramatically in rural areas by all socioeconomic regions between 2008 and 2015. This rise may be associated with a significant increase in the number of births occurring at secondary or higher-level (tertiary) hospitals, reflecting an increase in availability and accessibility of these services in these areas. The Chinese government had made a strong commitment to reduce maternal and child mortality to achieve the Millennium Development Goal targets by 2015.19 The main strategy was to promote hospital delivery, particularly in rural and poor areas, with largely financial support from the central government and partly from the provincial government.19 In the context of deepening China’s health system reform, the national plan of further strengthening the hospital delivery for rural women in 2009 was highlighted to provide the financial compensation for hospital delivery through rural health insurance (NCMS), the earmarked government fund and medical assistance programme for the poor women in order to reduce financial burden placed on the households.20 By 2014, hospital delivery in rural areas was almost universal. Across countries, the shift from community to hospital births is known to result in an increase in CS rates for medically indicated and non-medical reasons.21 In China, cross-sectional research has shown that while tertiary hospitals have the highest rates of CS for ambiguous indications (ie, non-reassuring fetal heart tracing, failure to progress), secondary-level hospitals report greater use of CS for maternal request.22

Implications for practice

Previous studies report maternal request for CS as a contributor to a rapid increase of CS rate in China,22–25 despite the validity of the concept being widely debated internationally. For women who reportedly prefer CS, the most common reasons for their preference are fear of labour pain, and safety for their baby and for themselves.11 In this context, family members (eg, husband or parent) also supported this choice to avoid an adverse event, especially in the context of one child in a family policy. Moreover, in our study, around one-third of nulliparous women reported self-request for CS in the era of the one-child policy. This proportion significantly decreased in urban areas over time, which may be associated with the shift from strict one child in a family in urban areas to universal two children, and promotion of vaginal births in hospital settings. This change did not occur in rural areas, which indicates the need for strengthening the quality of maternity care including service delivery and women’s experience in rural areas.

Efforts to promote vaginal birth in China included midwifery care training (eg, training more professional midwives, establishment of standardised evaluation scheme of midwifery practice, etc), pain relief for vaginal birth, and informing women about benefits and risks of different modes of delivery.26 Other studies, largely in big cities and tertiary hospitals, report woman-centred pregnancy and childbirth care, which includes provision of antenatal classes to shape women’s beliefs and confidence in childbirth, build connection and trust between doctors, midwives and women, as well as provide continuous support during labour and birth.15 23 27 At the same time, pharmacological and non-pharmacological options for labour pain management have become available. However, the midwifery workforce in China is insufficient. Quality of midwifery care can vary by hospitals, and urban–rural disparity in the number of midwives and training is anecdotally reported. Lack of support during labour, lack of pain relief and suboptimal birth environment were reported as the main reasons that rural women requested for CS.28 Hence, strengthening midwifery care to improve women’s experience on childbirth, particularly in rural areas, will be critical to optimise the use of CS in China.

Implications for further research

In the era of the two-child policy, only advanced maternal age and giving birth in a private hospital were positively associated with the use of CS by nulliparous women. Since 2013 onwards, the latest healthcare delivery reform in China encouraged competition between public and private hospitals, and set out the target of private hospitals sharing 20% of the market by 2015. The burgeoning of private hospital provision in China is driven by market forces, with providers’ charges unregulated unless contracted by the basic health insurance schemes.29 Studies in other countries report much higher CS rate in private hospitals due to profit-driven behaviour.30–32 It needs further study to have a better understanding on the use of CS in private hospitals in China in order to propose evidence-based recommendations for relevant policy development. In addition, we found that CS rate among multiparous women continued to increase over time, especially in urban areas, which may be associated with the increased number of women who underwent repeat CS. Based on the NMNMSS data, Liang and colleagues reported a high CS rate among multiparous women with a uterine scar and it was unchanged over time.9 In China, repeat CS is often suggested and accepted for women with a previous CS to mitigate the risk of uterine rupture or other adverse events, despite repeat CS having the similar risks. We can only speculate that there may be increasing referrals of women with a previous CS to high-level or specialty hospitals. The accessibility, functional referral and affordability of such services, as well as health outcomes, particularly for socioeconomic-vulnerable women, should be rigorously assessed in further research.

Conclusion

A rapid increase of CS rate in rural areas and the less developed western region contributed to the increase of CS rate in China over the past decade. The population policy shift, alongside facility policies to limit the use of unnecessary CS, are likely factors contributing to the reduction of CS in urban areas. Strategies at system, organisation and individual levels to mitigate unnecessary CSs should be continually strengthened, especially in rural areas and the western region. Improving midwifery care will be fundamental to ensure safety and positive childbirth experience in the era of the two-child policy in a family in China.