The distribution and accessibility of assisted reproductive technology clinics in mainland China from 2006 to 2018: a population-based retrospective study

Abstract Infertility is a global health problem that carries a high social and economic burden. Assisted reproductive technology (ART) has been developed in mainland China for over 30 years. We aimed to evaluate the accessibility and equity of distribution of ART facilities in mainland China from 2006 to 2018 and quantify the population with reduced geographic access to ART services. A retrospective study was conducted to describe the trend and analyse the equity of distribution of ART clinics in 2006, 2012, and 2018. The accessibility of ART clinics in mainland China increased significantly in the 12 years to 2018 (p ˂ 0.05). Eastern China had the most extensive coverage, followed by the central region, while the western region had the least coverage (p ˂ 0.05). The Gini coefficient for the distribution of ART clinics in 2018 was 0.213, indicating that the equity of distribution of ART clinics was relatively balanced over the country. However, at the end of 2018, there were still 354.9 million people (25.4% of the population) living in 148 cities without access to any ART clinics, which has spurred more targeted policies and cost-effective measures to improve the accessibility and availability of ART services in such areas.


Introduction
Infertility is a global health problem that carries a high social and economic burden, especially where access to care is lacking (ESHRE Task Force on Ethics & Law et al., 2008;Makuch et al., 2011;Quinn & Fujimoto, 2016;van Empel et al., 2010).Unequal distribution and improper allocation can result in the loss of resources and increase of health costs for the patients, the health system, and the whole society (Rezaei et al., 2016).Distance, cost, health literacy, ethnicity, and expertise are important factors that influence whether and how patients seek health care (Dyer et al., 2020; Ethics Committee of the American Society for Reproductive Medicine, 2015; Harris et al., 2017;Warner et al., 2015).As Assisted Reproductive Technology (ART) involves multiple consultative appointments, complex diagnosis, invasive therapies and partner-based care, disparity of burden on patients with infertility could be greatly magnified.The World Health Organisation (WHO) and many fertility organisations have devoted great efforts to improving access to infertility care and reducing inequalities within and between countries (American Society for Reproductive Medicine, 2015; ESHRE Task Force on Ethics & Law et al., 2008; Ethics Committee of the American Society for Reproductive Medicine, 2015).Recent publications by the American Society of Reproductive Medicine (ASRM) and the Centres for Disease Control and Prevention (CDC) advocate eliminating or reducing the disparities in access to infertility care, especially disparities between different ethnicities, socioeconomic classes, and regions (Ethics Committee of the American Society for Reproductive Medicine, 2015).China has a vast territory, and inequality of health care exists not only between provinces or regions, but also exists at the intra-province level (Yin et al., 2018).
ART has been developed in mainland China for over 30 years since the first test tube baby was born in Beijing in 1988, and is being applied more and more widely (Qiao & Feng, 2014).The first regulation of ART in mainland China was promulgated in February 2001, stipulating those medical institutions that carry out ART treatment should be licenced and monitored by the National Health Commission (NHC) (2003).ART was classified as artificial insemination with husband's semen (AIH), artificial insemination with donor semen (AID), in vitro fertilisation and embryo transfer (IVF-ET), intracytoplasmic sperm injection (ICSI), and preimplantation genetic testing (PGT).In 2007, the NHC transferred the right of licencing to local Health Commissions.To strengthen the management of ART, the NHC has issued a series of supplementary regulations and documents since 2013, including supplementary provisions on approval requirements and guiding principles for planning the allocation of ART clinics for the period from 2015 to 2020 (Bai et al., 2020;National Health Commission, 2015a, 2015b).The number of clinics in each province is defined by the government.One of the optional criteria is that one ART clinic should be set up for every 3 million permanent residents in a province.
In this context, this study attempts to summarise and describe the allocation and development of ART facilities in mainland China in recent years, evaluate the equity of accessibility of ART facilities in different periods, and identify and quantify the populations with reduced access to ART services.

Data sources
The NHC has published the list of approved ART clinics in mainland China six times : in 2004, 2006, 2007, 2012, 2016and 2018(National Health Commission, 2019)

Data analysis
A retrospective study was conducted to describe the trend and analyse the equity of distribution of ART clinics in 2006ART clinics in , 2012ART clinics in , and 2018.Given data availability, we used the city as our statistical unit.There were 333 prefecture-level cities and 4 province-level municipalities (Beijing, Tianjin, Shanghai, and Chongqing) in mainland China (Ministry of Civil Affairs of the People's Republic of China, 2019).The distribution of ART clinics was summarised into two levels: at national and at regional level using three regions: eastern, central, and western (Figure 1).
Regional choice was defined in five levels as a city with: (i) no ART clinic; (ii) a single artificial insemination (AI) clinic; (iii) multiple AI clinics but without IVF clinic; (iv) a single IVF clinic; or (v) multiple IVF clinics.In 2018, there were also four clinics allocated in counties directly under the provincial government.These were excluded in our analysis.We described accessibility by calculating the proportion of population with each regional choice.
The Gini coefficient is widely used to measure the balance of the distribution of health resources.It was calculated to analyse the equity in distribution of ART clinics in the whole country and in each region.The Gini coefficient has a range from 0 to 1.The closer it is to 1, the greater the inequity in distribution of resources it represents: (i) 0.2 absolute equality; (ii) 0.2-0.35relatively balanced; (iii) 0.35-0.5 comparatively unequal distribution; (iv) 0.5-0.7 very uneven, and so on (Brown, 1994;Lu & Zeng, 2018;Mobaraki et al., 2013;Wang et al., 2020).The equation for calculating the Gini coefficient is as follows: where Y i is the cumulative percentage of ART/IVF clinics in each province, X i is the cumulative percentage of population (sorted by variable) of each province, and n is the number of provinces.

Statistical methods
Regional choice of each city was plotted graphically using ArcMap 10.2 bound by city regions.Coverage rates among groups were compared using the Chi-Square test or Cochran-Armitage Test for Trend.Due to heterogeneity of variance, medians among groups were compared using Kruskal-Wallis rank sum test.These two analyses were performed by SPSS 22.The Gini coefficients were calculated by Excel 2016.
All data were taken from the publicly available sources.These were secondary data without any personal information, so informed consent was not needed.

General characteristics
In mainland China, ART includes AI, IVF and various derived techniques, however the data about the clinics which only offer AI services was lacking in 2006.To measure the precise trend of the ART clinics development, we compared the growth rate of IVF clinics from 2006 to 2018 and the growth rate of all ART clinics from 2012 to 2018 respectively.There were 85 clinics that offered IVF technology in mainland China in 2006China in , 217 in 2012China in , and 375 in 2018 respectively.The growth rate (155.3%) in the number of IVF clinics in the first stage (2006)(2007)(2008)(2009)(2010)(2011)(2012) was faster than the equivalent growth rate (72.8%) in the second stage (2012)(2013)(2014)(2015)(2016)(2017)(2018).The national total of ART clinics in mainland China increased from 356 in 2012 to 498 in 2018, with an increase rate of 39.9% (Table 1).
Recently, the NHC has announced ART Allocation Planning Guideline for the period from 2021 to 2025, in which the data of cycles and populations of each province in 2018 were provided.Cycles per million  population (C/M) in mainland China in 2018 was 774.9.C/M in eastern, central and western China were 1066.4,580.4 and 555.7 respectively (Table 1).

Accessibility analysis
In 2006, 40 (11.9%)cities had IVF clinics, covering a population of 291.2 million (22.5% of the national population).By the end of 2018, 142 cities (42.1% of the number of cities nationally) had IVF clinics, covering a population of 870.2 million (62.3% of the national population).A total of 189 cities (56.1% of the number of cities) had ART clinics, covering 1.045 billion people (74.8% of the national population).The coverage rates calculated by city or by population in mainland China and each region increased significantly in the 12 years (p ˂ 0.05), except that there was no statistical difference in the coverage rate of ART clinics calculated by cities in central China between 2012 and 2018 (Table 2).Overall, the eastern region had the most extensive coverage, followed by the central region, and the western region, which had the least coverage (Table 2, Figure 2).The differences in distribution of ART and IVF clinics in the three regions were statistically significant (p ˂ 0.05).In 2006, 2012 and 2018, the median population of cities with multiple IVF clinics was 9.8 million, 7.7 million and 7.3 million respectively, which suggests a significant improvement in accessibility.It shows that better ART services were distributed in the cities with larger population (Table 3).

Equity analysis
The Gini coefficients for the distribution of IVF clinics in mainland China in 2006, 2012, and 2018 based on population index were 0.389, 0.275 and 0.267 respectively (Figure 3).The Gini coefficients for the distribution of ART clinics in mainland China in 2012, and

Discussion
In this study we investigated trends of allocation of ART clinics allocation over the previous 12 years in mainland China.We also evaluated the equity of distribution of ART clinics.This analysis indicates that the number and accessibility of ART clinics in mainland China increased significantly from 2006 to 2018.The Gini coefficients showed that from 2006 to 2018, the equity of distribution of ART clinics improved greatly, reaching relative balance overall.The distribution of IVF clinics in eastern China and ART clinics in eastern China and central China reached absolute equality.
By 2018, the total number of ART clinics in mainland China was 498, which is close to the total number of ART clinics in the United States (502 in 2016), much more than the number in United Kingdom (82 in 2016), and less than the number in Japan (622 in 2018) (European IVF-monitoring Consortium (EIM) for the European Society of Human Reproduction and Embryology (ESHRE) et al., 2020;Ishihara et al., 2021;Sunderam et al., 2018).Though ART services developed a lot in the 12 years, if it is compared to the population size, the ART service in 2018 was still at a lower level than other well-developed countries.Recently, the NHC of China has issued new round ART Allocation Planning Guideline for the period from 2021 to 2025, in which three calculation methods were provided to the provincial governments to

Implications for practice and policies making
Despite ART becoming a mainstream medical intervention for infertility, the accessibility of ART clinics in western China was still at a low level in 2018, which implies several factors.On the one hand, previous analysis has indicated that ART clinics were more likely concentrated in cities with a large population, which is consistent with ART clinic distribution in the United States (Nangia et al., 2010;Yin et al., 2018).Many provinces in western China have the geographical characteristics of vast territory and sparse population, and much of the territory is at high altitude.Thus, access to clinics is difficult for many people in those provinces.On the other hand, the cost to establish an ART clinic is high.The high investment needed for facility construction, advanced equipment, professional technology, and human resources could be intangible impediments to efforts by local government to establish ART clinics.In China, infertility treatments are not yet subsidised through the national health insurance scheme.Residents' income determines whether they are willing or able to make full use of ART.A previous study in Australia showed that women in disadvantaged social-economic classes had a 6-16% reduction  in access to ART treatment, and women living in regional and remote areas had a 12% reduction.If the out-of-pocket cost of ART treatment was increased by one third there would be a 21-25% decrease in access to treatment (Chambers et al., 2013;Harris et al., 2016).The overall economic development of western China is lagging behind that of the eastern and central regions and has resulted in a reduction of both the supply and utilisation of ART services.
In future rounds of regional allocation, the geographical and social-economic characteristics of western China should be taken into consideration.Basic infertility diagnosis and treatment technologies, such as prevention and screening of impaired fecundity, reproductive health education, and cost-effective treatments should be given priority.In addition, local monitoring and tele-medicine can be used to relieve some of the cost burden for patients in areas without advanced ART services.In the second health care system reform since 2010, the Chinese government has attempted to resolve the disparity of medical resources by grouping medical institutes of different levels into a regional medical consortium (Li et al., 2017).Establishing and promoting dual referral systems and appointments, data-sharing between hospitals and regional information systems, and remote consultation services are part of the reform goals (Lei et al., 2017;Liang et al., 2019).These changes will offer opportunities and raise a promising prospect for improving the availability and accessibility of ART treatment.
This study has certain limitations.First, it was assumed that if one or more ART clinics were in a city, all residents in the city had access to ART services.Some factors, such as the great distance or high cost that discourage people from utilising fertility health care were not considered.From this perspective, the estimation of 179.5 million reproductive-age population with no access to ART services may be lower than reality.On the other hand, some advanced ART clinics located near the border of provinces could attract patients from outside the province (Nangia et al., 2010).That might cause overestimation of people with limited accessibility.
In addition, this study only analyses the equality of the distribution of ART facilities in mainland China from the service providing perspective.Factors which influence the utilisation of infertility service and how the ART service are actually utilised are not involved.To better recognise and furthermore reduce the inequity in the provision of ART services and lower the barriers to utilisation of those services, more endeavours in related research will be needed in future, which is the focus of our future direction.This study has tried to offer a glimpse of the development and distribution of ART clinics in China based on the available empirical evidence.It is hoped that these results can serve as an impartial and objective benchmark for future policy making and assessment in China.
In conclusion, this study shows that the accessibility of ART clinics in mainland China has increased significantly in the last 12 years.The equity of ART clinics distribution is now relatively balanced over the country based on the population index.However, underserved areas do exist, mostly in western China.Besides properly increasing the number of ART clinics, more cost-effective measures to improve the accessibility and availability of infertility prevention and health care should be given priority in underserved areas.

Figure 1 .
Figure 1.Number of cities in each province and regional division in china.
2018 based on population index were 0.222 and 0.213 respectively.From 2006 to 2012, the Gini coefficient for the distribution of IVF clinics in the eastern and central regions declined greatly.As for the distribution of ART clinics in the three regions in 2012 and 2018, the Gini coefficient changed only slightly.

Figure 2 .
Figure 2. Map of China by regional access to ART in 2006, 2012 and 2018.The legend defines areas where cities have no ART clinies, only one AI clinic, multiple Al clinics without IVF, one IVF clinics.Each city is outlined in light grey.Each province is outlined in dark grey.

Figure 3 .
Figure 3. Gini coefficients for distribution of ART/IVF clinics in China based on population.

Table 1 .
Number and growth rates of ART/IVF clinics in 2006, 2012 and 2018, with the cycles in 2018.These are calculated based on the data from Human Assisted Reproductive Technology Allocation Planning Guideline (version 2021) (National HealthCommission, 2021).Cycles herein refer to the total number of fresh IVF and ICSI oocyte retrieval cycles, frozen embryo transfer (FET) thawing cycles and PGT diagnosis cycles.
Table 3, a total of 576.3 million people in 73 cities had access to multiple IVF clinics in 2018, compared to 208.0 million people in 23 cities in 2006.A total of 171.4 million people in 47 cities had no access to an IVF clinic, but had access to at least one AI clinic.By 2018, there were still 354.9(25.4%) million people living in 148 cities without any ART clinic.Using the 2010 Population Census Data of China, our calculations indicate that almost 179.5 million reproductive-age (ages 20-49 years) people (91.1 million male and 88.3 million female) had no access to any ART services.

Table 2 .
The city and population coverage rates ofART/IVF clinics in 2006, 2012 and 2018.To compare the coverage rates of ART/IVF clinics in 2006, 2012, and 2018 using Cochran-Armitage Test for Trend.p-Value less than 0.05 is statistically significant.ÃÃ To compare the coverage rates of ART/IVF clinics in Eastern, Central, and Western China using Chi-Square test.p-Value less than 0.05 is statistically significant.ÃÃÃ A list of clinics which only carried out AIH treatment was not available in 2006.ÃÃÃÃ 0 cells have an expected count of less than 5.

Table 3 .
ART health care availability for the Chinese mainland population in 2006, 2012 and 2018.To compare the medians of population among different levels of ART/IVF clinic accessibility in 2006, 2012, and 2018 using Kruskal-Wallis rank sum test.ÃÃ List of clinics which only carried out AIH treatment were not available in 2006.