Cumulative live birth rate according to the number of receiving governmental subsidies for assisted reproductive technology in Saitama Prefecture, Japan: A retrospective study using individual data for governmental subsidies

Abstract Purpose We investigated the cumulative live birth rate (CLBR) in women receiving governmental subsidies for assisted reproductive technology (ART) in Saitama Prefecture, Japan. Methods Women who applied for subsidies from Saitama Prefectural Government for the first time in 2016 were enrolled and followed up until the end of 2017. Treatment information, including live birth, was obtained from the Japanese ART registry by linking it with unique identification numbers for treatment. Patients’ factors associated with having a live birth were investigated. Results Of 1,072 women (2,513 applications), 495 (46.2%) had a live birth with 8 (1.6%) twin pregnancies. The CLBR over six subsidized cycles was 53.7% for women aged <40 years, and 17.2% over three subsidized cycles for women 40‐42 years; highest among women <35 years (58.4%), followed by those aged 35‐39 years (49.3%). Multivariate analysis revealed patient age as the only independent factor for having a live birth. Conclusions The CLBR of women receiving subsidies for ART was greatest in women aged <35 years. Effective policies for promoting ART among younger couples who seek infertility treatment are essential.

despite a rapidly decreasing trend of the total numbers of babies born in Japan. 2 One of the main reasons for the increased number of ART cycles in Japan is attributable to the advanced age for women seeking this option. As is common in many developed countries, 3 women tend to delay child bearing in Japan, 4 and face an increased risk of infertility related to advanced age; from the report of the JSOG, 41.8% of registered ART cycles were those for women in their 40s, 1 which is extremely high compared with other countries. 5 Although the health insurance system in Japan provides universal coverage, 6 ART treatment is an exception, and patients must pay for ART treatment out of their own pockets. However, the Japanese government currently reimburses a part of the ART treatment cost for six attempts for women aged <40 years, and for three attempts for women aged 40-42 years since 2014. There was an upper limit of 7 300 000 JPY (approximately 70 200 USD using a 2021 exchange rate of 1 USD = 104 JPY) per annual couple's income for receiving the subsidies, but this policy was modified in January 2021 because of the recent stagnation in the Japanese total fertility rate. 7 The newly introduced elimination of the income limit is expected to further increase the number of ART treatment cycles. However, to date, the live birth impact of governmental subsidies for ART treatment has not been fully evaluated. 8 More specifically, among those receiving subsidies, little is known about who had a live birth in terms of age groups, infertility reasons, income levels, or the number of subsidized cycles. Such evaluation is essential for drafting an effective policy for the utilization of ART treatment in Japan. However, in Japan, nationwide individual-based data on the governmental subsidies for ART is not available, partly because the Ministry of Health Labour and Welfare has not collected such data and partly because the governmental subsidies were mainly managed by local municipalities including a core city (populations >200 000), an ordinancedesignated city, and otherwise prefectural government. Therefore, we aim to investigate CLBR according to the number of receiving governmental subsidies in women stratified by different age groups using individual data managed by Saitama Prefectural Government.

| MATERIAL S AND ME THODS
This is a retrospective study using individual data based on the application of governmental subsidies for ART managed by Saitama Prefectural Government. We applied for access to individual data for ART subsidies in anonymous form to Saitama Prefectural Government during 2016 and 2017. In Saitama prefecture, governmental subsidies were managed by core cities with populations >200 000 (Koshigaya, Kawagoe, and Kawaguchi), and an ordinancedesignated city (Saitama city) and otherwise Saitama Prefectural Government, and the same applied to the individual data related to them. The population and number of households covered by this study in 2016 were 5 315 797 and 2 978 871, respectively: being 73.1% and 72.5% of the overall Saitama Prefecture because the above cities were excluded. Kawaguchi was designated as a core city in 2018, and applicants from the city before 2018 were included in this study. Subsidies for ART are eligible for couples in which the woman is aged less than 43 years and with an annual couple's income of less than 7 300 000 JPY. This study was approved by the institutional review board of Saitama Medical University (Approval number, 904; September 2019) and the ethics committee of the JSOG (Approval number 2020-2; June 2020). After these approvals, the Saitama Prefectural Government provided us with access to data without any identifying information.
Individual data for application used for the analysis included the number of applications, patients' age, and the husband's and wife's annual income. Treatment types were classified as follows: A, cycles with fresh embryo transfer (ET); B, Cycle with freeze-all and a subsequent frozen-thawed embryo transfer (FET); C, FET cycles; D, cancelation because of patients' health problems; E, cancelation because of failed fertilization or embryo development; and F, cancelation because of failed oocyte collection. We also used the fertilization methods for fresh ET cycles including in vitro fertilization (IVF), intracytoplasmic sperm injection (ICSI), and the results in terms of conception from the application information. Husband's and wife's annual income was based on a proof of earnings for prefectural tax or a taxation certificate. Women who reported no income were defined as housewives. The data also included unique identification numbers for the designated treatment cycle based on the Japanese ART registry in which almost all treatment cycles conducted in Japan are registered. This study linked the individual data for application and the Japanese ART registry to confirm a live birth status. From the Japanese ART registry, information on infertility diagnosis, fresh/FET status, ovarian stimulation protocols for fresh cycles, fertilization methods (IVF/ICSI), number of collected oocytes, number of embryos transferred, and number of embryos cryopreserved and pregnancy outcomes including live birth status was used for the analysis.
A flow diagram of this study is shown in Figure 1 Using these data, we have calculated the cumulative live birth rate (CLBR) according to the number of applications for subsidies.
The definition of CLBR was the first live birth per individual patient.
We took a conservative approach to calculate this, assuming that none of the women who discontinued making applications would have had a live birth. Thus, the numerator was the number of live births divided by the total number of women included. Once a woman had achieved her first liveborn baby, they did not contribute to any further CLBRs for our analysis. CLBRs were calculated for three age groups based on the first application for a subsidy in 2016 (<35, 35-39, and 40-42 years). For women aged 40-42, governmental subsides were only allowed for three ART cycles, and CLBRs were calculated up to this level for all.

| Statistical analysis
First, application information was analyzed for the included and excluded samples among the eligible sample (n = 1807) to evaluate the pattern of inclusion. Then, baseline characteristics were analyzed according to different age groups using chi-squared or Fisher's exact tests or one-way analysis of variance as appropriate. Among the analyzed sample, information on fresh cycles at the first application was available for 980 women, and the treatment information for fresh cycles was analyzed according to the different age groups. Finally, patient factors associated with having a live birth were investigated.
Because the outcomes were not rare (ie, >10%), we directly estimated relative risks (RRs) of patient factors for having a live birth using a generalized linear model for a binomial family with log-link function instead of calculating odds ratios using logistic regression modeling. 9 In the multivariable model, age at the time of first application in 2016, infertility diagnosis, quartiles of annual couple's income, and housewife status were included. All analyses were conducted using the STATA MP statistical package, version 16.0 (Stata).
A two-tailed P value of <.05 was considered statistically significant. Baseline characteristics for the analyzed sample stratified by the women's age are shown in Table 2. For the treatment type at first application, the proportion of freeze-all and a subsequent FET cycles was higher in the younger age group, while the proportions of canceled cycles were higher in the older age group with statistical significance (P < .001). For infertility diagnosis, the proportions of tubal factor and male factor infertility diagnoses were significantly higher in the younger age group (P = .005 for tubal factor and P = .04 for male factor, respectively), while other infertility diagnoses were more prevalent in the older age group.

| RE SULTS
The CLBR according to the number of applications stratified by women's age is shown in Figure 2. Of 1,072 women included, 495 (46.2%) actually had a live birth, of which eight (1.6%) were twin pregnancies. 339 live births (68.5%) were derived from FET cycles, and there was no significant association between treatment type (A-C) and patients' age in live birth cases (P = .71). The CLBR over six ART subsidies was 53.7% for women aged <40 years and 17.2% over three subsidies for women aged 40-42 years; it was highest among women aged <35 years (58.4%), followed by those in the 35-39 years group (49.3%). The CLBR increased up to the fourth or fifth application but plateaued gradually between that and the sixth application. Detailed numbers of applicants and live births according to the number of governmental subsidies are shown in Table S1.
Treatment information for fresh cycles at the first application stratified by age group is shown in Table 3. For ovarian stimulation, natural cycles were selected for 14% of women across all age groups, while mild ovarian stimulation using clomiphene citrate (CC) alone was more often F I G U R E 1 Flow diagram of sample selection selected in the older group. GnRH agonist and antagonist protocols accounted for almost half of the cycles for women aged <35 years, but the proportions decreased with age. The mean number of oocytes collected was highest in those aged <35 years (mean 8.0, SD ± 7.2), while that in the 40-42 years group was 4.2 (SD ± 5.2). Fresh ET was conducted in 44% to 48% of all age groups, and single embryo transfer (SET) was conducted in 98.8% of women aged <35 years, in 90.3% of those aged 35-39 years, and in 80.0% of those aged 40-42 (P < .001). The mean number of frozen embryos was the highest in those aged <35 years (mean 2.5, SD ± 3.1), while it was the lowest (mean 1.0, SD ± 1.7) in those aged 40-42 years with statistical significance (P < .001).
The RRs and 95% CIs of patient characteristics for having a live birth are shown in

| D ISCUSS I ON
In this retrospective study using individual data for women receiving governmental subsidies based on their first application in 2016, linked with the Japanese ART registry in which detailed treatment information was included, we found that the overall CLBR across six sets of subsidies for women aged <40 years was 53.7%; highest among women aged <35 years (58.4%), followed by those aged 35-39 years (49.3%). For women aged 40-42 years who were allowed three sets of subsidies, the CLBR was 17.2%. Multivariate analysis revealed only patient age as an independent factor associated with having a live birth. To date, this is the first study using individual treatment information to evaluate the CLBR according to the number of governmental subsidies for ART cycles in Japan.   to infertility care, 11 and Japan had the highest worldwide (3,212 cycles/million), which was ~500 times higher than Senegal, the country with the lowest rate (6 cycles/million). Importantly, the mean age of women receiving ART was extremely high in Japan: The proportion of registered cycles applied to women aged ≥40 years was 41.8% in 2018. This compares with the following other developed countries: 28.3% in Australia, 22.1% in Germany, and 23.9% in the United States according to the ICMART report. 5 Because ART using donor oocytes/ embryos is almost never practiced in Japan, older women have to continue treatment multiple times to achieve a live birth compared with younger women, so advanced patient age has been the major reason for the increased numbers of ART treatment cycles in Japan.
To improve the live birth impact of governmental subsidies for ART, age limitations for subsidies and incentives for younger women were introduced in 2016; subsidies were only available for women aged <43 years, and patients aged 40-42 y can only apply for three attempts, while patients aged <40 years can apply for six attempts. In fact, the CLBR in our analysis was the highest in women aged <35 years (58.4%).
Considering that patient age was the only significant background factor for having a live birth in our study, incentives for younger couples would be a reasonable strategy for effective subsidies.
One of the other clinical outcomes in this study was a multiple pregnancy rate of only 1.6% (eight sets of twins), which is exceptionally low compared with other developed countries. 5 This low prevalence is largely attributed to the high SET rate of 82%-83% in 2018 12 in Japan following JSOG recommendations in 2008. 13 Moreover, the high accessibility of Japanese to ART treatment might also have played a role. It has been hypothesized that countries with restricted accessibility to ART, such as those with underinsurance or very limited health insurance coverage, tend to choose aggressive treatments such as ovarian hyperstimulation and multiple embryo transfer, 14 which resulted in very high multiple pregnancy rates. 15 In Japan, the ease of access to ART treatment might help decrease the demand of patients for more aggressive treatments. To reduce the high burden of healthcare costs of caring for multiple-birth infants after ART as well as maternal risks for pregnancy complications, [16][17][18] several countries (eg, Belgium, New Zealand, and Turkey) and states (eg, Quebec in Canada and Connecticut in the United States) have introduced health insurance coverage for ART linked to SET policy, and have successfully achieved significant reductions both in multiple pregnancy rates and in related healthcare costs. [19][20][21][22] In this study, nearly 60% of younger patients (<35 years) had a live birth after at least 1 year. To date, evaluation of the Japanese governmental subsidies, in terms of clinical practice, and outcomes, such as the CLBR, have not been fully investigated. Because the Japanese ART registry only includes cycle-specific information and cannot distinguish between individual patients receiving multiple treatment, only live birth rates per total number of treatment cycles with or without using subsidies were available. Thus, the CLBR using our data would be informative for infertile couples to decide on the implementation of ART using governmental subsidies.
This is the first study to investigate the impact of governmental subsidies for ART on the CLBR in Japan using individual data for governmental ART subsidies linked with the Japanese national ART registry. However, there were several limitations. First, we linked individual data with the Japanese ART registry using unique IDs for treatment cycles, but a significant proportion of these (39.0%) lacked this information. This might have introduced selection bias for the sample selected. Characteristics based on application information stratified by included and excluded status demonstrated that treatment type, the mean total number of subsidies, and mean annual couple's income were significantly different between the included and excluded samples ( were not counted in this study due to the methodological limitation.
Third, although we linked the data with the Japanese ART registry, the study still lacked important confounding information such as any history of previous children arising from ART treatment, 23 body mass index, 24 and duration of infertility. 25 These unmeasured confounders would invariably affect the multivariable analysis.
Although the Japanese government has recently indicated it plans to cover infertility treatment by social health insurance by 2022, the establishment of sustainable healthcare systems is essential. For example, in Quebec, Canada, it was widely reported that the abovementioned universal coverage of IVF in conjunction with a SET policy meant that the access to IVF was dramatically improved, 26 which resulted in a significant decline in the multiple pregnancy rate. 22 However, in 2015, the system was dismantled because of costs to the healthcare system. 19 Thus, for a sustainable system, clear indications and appropriate restrictions for the use of ART under health insurance coverage are necessary. In addition, continuous monitoring of ART utilization and outcome measurements will be indispensable.

| CON CLUS IONS
In conclusion, this study, using individual data for governmental ART subsidies and linked with the Japanese national ART TA B L E 3 Treatment information for fresh cycles at first application stratified by age at the first application for ART subsidy (n = 980) a registry, demonstrated that the overall CLBR across six subsidies was 53.7%, in which women aged <35 years had the highest rate (58.4%), while 49.3% of the women aged 35-39 years actually had a live birth. For women aged 40-42 years for whom only three subsidies were allowed, the CLBR was 17.2%. Given that age was the only significant factor associated with CLBR, policies for promoting ART among younger couples who seek infertility treatment in Japan are essential.

ACK N OWLED G M ENTS
We thank Saitama Prefectural Government for providing anonymous individual data for governmental subsidies. Further, we thank all of the registered facilities for their cooperation in providing cyclespecific information into the ART registry. This study was supported by a Health and Labor Sciences Research Grant (H30-Sukoyaka-Ippan-002). We thank James Cummins, PhD, from Edanz (https:// en-autho r-servi ces.edanz.com/ac) for editing the English text of a draft of this manuscript.

CO N FLI C T O F I NTE R E S T
Dr Osamu Ishihara has received an honorarium from Ferring Pharmaceuticals.

E TH I C S A PPROVA L
This study was approved by the institutional review board of Saitama Medical University (Approval number, 904; September 2019) and the ethics committee of the JSOG (Approval number 2020-2; June 2020).

H U M A N R I G HT S S TATE M E NT A N D I N FO R M E D CO N S E NT
All procedures were performed in accordance with the ethical standards of the relevant committees on human experimentation (institutional and national) and the Helsinki Declaration of 1964 and its later amendments.

A N I M A L R I G HTS
This report does not contain any studies performed by any of the authors that included animal participants.