Assessment of patient preferences for assisted reproductive technology in China: a discrete choice experiment

Abstract

Objectives

Given China’s low fertility rate, assisted reproductive technology (ART) can be used assist infertile patients in having children. This study aimed to analyse patients’ preferences for ART and to determine the relative importance (RI) and willingness to pay (WTP) of key attributes.

Design

We identified six attributes of ART and used a D-efficient design to generate choice sets for conducting a discrete choice experiment. Patients were asked to choose between two scenarios that differed in participation in treatment decision-making (TDM), clinical pregnancy rate, live birth rate, risk of maternal and neonatal complications, and out-of-pocket cost.

Setting

Jiangsu province, China. The anonymous survey was carried out between December 2022 and February 2023.

Participants

Female patients aged 20–45 years, with low fertility or experience of ART treatment. We recruited 465 participants.

Outcomes measures

Patient-reported preferences for each attribute were estimated using a mixed logit model. The latent class model was also used to investigate preference heterogeneity.

Results

All attributes were associated with patient preferences. Patients considered the live birth rate as the most important attribute (RI=29.05%), followed by participation in TDM (RI=21.91%). The latent class model revealed two distinct classes named ‘outcome driven’ and ‘cost driven’. Preferences varied according to their age, monthly household income and location.

Conclusions

This study investigated the preferences of infertile patients when seeking medical assistance for infertility. The study outcomes can contribute to evidence-based counselling and shared decision-making and provide an empirical basis for creating and implementing future policies.

Correspondence to Dr Xin Li; [email protected] ; Dr Lingli Zhang; [email protected]

STRENGTHS AND LIMITATIONS OF THIS STUDY

• The experimental design and data analysis were conducted following the International Society for Pharmacoeconomics and Outcomes Research Conjoint Analysis Task Forces and the latest DIRECT Checklist and the choice experiments were tested and refined through a piloting process.

• Efficient design and blocking design were used in the questionnaire design to improve statistical efficiency and patient response.

• The participants were recruited from Jiangsu Province and comprised women who were either infertile or undergoing assisted reproductive treatment; results may not be applicable to other regions and populations.

• While we verified the internal validity of the discrete choice experiment through consistency tests and an exit option design, the external validity could not be verified.

Introduction

In recent years, China has witnessed a sustained decline in fertility and population growth rates. According to the 2023 China Statistical Yearbook, the national birth rate decreased to 6.39‰, while the natural growth rate of the domestic population was recorded at −1.48‰. This marks the second period of negative population growth in China in nearly 61 years. The increasing incidence of infertility has contributed to the decline in fertility rates. Globally, it is estimated that 48 million couples and 186 million individuals are affected by infertility.¹ According to statistical data, the infertility rate in China is expected to reach 18.2% in 2023.²

The introduction and application of assisted reproductive technology (ART) have undoubtedly provided significant benefits for individuals experiencing infertility. ART refers to the artificial operation of gametes, zygotes and embryos through medically assisted techniques to enable infertile couples to conceive a child.³ Nevertheless, ART comes with certain risks, including high expenses, potential complications and the possibility of treatment failure. The costs of one single ART cycle typically cover ovulation drugs, surgery, embryo freezing, preoperative examination and other associate costs.⁴ The financial burden of this treatment ranges from $4453 to $29 686,² which is equivalent to 0.6–4 times the per capita disposable income of residents in Jiangsu in 2022 ($7401). The extended duration and significant cost of ART contribute to a considerable economic strain on patients, their families and society. The cost of treatment has been shown to influence patients’ decisions regarding their treatment options. High costs still keep many infertile families out of treatment. In recent years, the State has continued to improve the maternity insurance system and has actively promoted the inclusion of assisted reproduction services in health insurance coverage. As of now, 27 provinces in China have successively included ART in medical insurance reimbursement. Incorporating cost attributes into discrete choice experiment (DCE) studies and estimating patients’ willingness to pay (WTP) can provide valuable insights for the development of future health insurance reimbursement policies.

High-quality healthcare should be customised to meet the specific needs of patients. Patients often exhibit a diverse range of preferences and may hold perspectives that differ from those of their physicians, posing a challenge for both clinicians and patients to reach a consensus. Understanding patient preferences is crucial for enhancing service quality and improving pregnancy outcomes. This study aims to assist physicians in comprehending patients’ needs and values, thereby enhancing patient satisfaction and adherence to treatment. Additionally, it will serve as a foundational basis for the formulation of policies and the allocation of resources that are more aligned with patients’ expectations. The selection of ART is a complex decision-making process influenced by a multitude of factors, with individual preferences exhibiting considerable variability. To date, only one study in China has employed rating-based conjoint analysis to examine the preferences of fertility care.⁵ The findings indicated that patients prioritised their doctor’s attitude above other factors, with success rate being the second most valued aspect. However, this study did not incorporate a WTP measurement, and the use of a scored questionnaire format may have increased the cognitive burden of trade-offs and subjectivity of respondents’ responses, which could have led to bias. The majority of the remaining relevant preference studies have been conducted in Western countries, such as the Netherlands and Australia, with limited extrapolation of the findings.^5–7 Notably, a large cross-national DCE study included participants from the USA, the UK, China and others.⁵ The results revealed that in the Chinese sample, patients with low fertility or experience of ART did not have a significant preference for cost, in contrast to other countries where ART is subsidised. Recently China has continued to enhance its maternity insurance system and promote the inclusion of assisted reproduction service programmes in healthcare payments. This undermines the significance of the previous studies. The significance of incorporating cost attributes into DCE studies and estimating patients’ WTP could inform future health insurance reimbursement policy development. With this empirical study, we aimed to investigate the preferences and WTP for ART among infertile Chinese patients.

Methods

Survey design

DCE has gained wide usage in healthcare for studying preference selection in recent years.⁶ It employs mathematical models to evaluate participants’ preferences for specific goods or services, WTP and predicted uptake rate.⁷ A set of attributes and levels can describe the profile of ART characteristics. According to random utility theory, respondents will typically select the option with the highest utility from the presented alternatives. The design and analysis adhered to the checklist and report of the International Society for Pharmacoeconomics and Outcomes Research and the latest DIscrete choice experiment REporting ChecklisT.^{6 8–10}ed We used a framework based on an extensive literature review, which identified three major categories of attributes: outcome attributes, process attributes and cost attributes.⁹ The attribute level determinations in this paper are based on the literature review and best-worst scaling (BWS) study primarily from the patient’s perspective. We combined quantitative and qualitative attributes to summarise the levels of attributes applied to ART at both national and international levels by DCE (details in online supplemental table S1). Based on the attribute pool, we conducted a (case 1) BWS study of 150 patients to determine the six attributes that patients value most in the treatment process. The results are shown in online supplemental table S2. The study ultimately identified six attributes, as detailed in table 1.

Table 1

Attributes and levels for DCE choice set

Attribute	Level
Participation in TDM	Not participated in TDM
	Partially participated in TDM
	Fully participated in TDM
Clinical pregnancy rate	20%, 40%, 60%
Live birth rate	15%, 35%, 55%
Risk of neonatal complications	5%, 10%, 15%
Risk of maternal complications	0.1%, 2.5%, 5%
Out-of-pocket costs per cycle (¥)	3000, 30 000, 50 000

US$1=¥6.737; https://www.oecd.org/en/data/indicators/exchange-rates.html (2022).

DCE, discrete choice experiment; TDM, treatment decision-making.

The questionnaire consists of a set of choices that requires a detailed description of its attributes and levels to ensure that the respondents fully comprehend the choice tasks (details in online supplemental table S3). To increase feasibility, the preference questionnaire used a D-efficient design, which was developed by Ngene, V.1.4 (www.choice-metrics.com). To achieve more stable parameter estimation results with greater freedom in estimation, 24 choice sets can be created.⁶ The Cond command was employed in the design to prevent illogical scenarios where the live birth rate exceeds the pregnancy rate. A blocking design can then be implemented to reduce the cognitive burden on patients. Among these sets, the questions for consistency testing were selected for repetition. Three versions of the questionnaire were generated, each containing nine choice sets. In addition, patients have the option to decline a health service regardless of its attributes and level. The unforced task provided an opt-out option, requiring respondents to choose between the selected profile from the forced task and the opt-out option.^{10 11} Furthermore, a questionnaire was employed to collect basic data. After consultation with experts (one associate professor of medicine and one nurse from the Reproductive Centre) to ensure the practicability of the study design under the current medical and policy framework, a pilot study was initiated. We carried out the study with a sample of 40 patients at Nanjing Drum Tower Hospital before the formal study commenced to validate its feasibility, acceptability and reliability. The results indicated that at least one level of each attribute was statistically significant (see online supplemental table S4). Respondents generally understood these attributes and levels. Feedback received and prior information were used to refine the study. An example of the final choice set is shown in online supplemental figure S1. The presentation of the questionnaire (block 1) can be found at the end of the online supplemental file.

Patient and public involvement

Patients participated in the attribute level determination phase and the pilot study. These helped to better complete our DCE study design.

Study population and data collection

Currently, ART has developed on a technological scale with 622 medical institutions authorised to provide service in China. Many general hospitals are establishing centres for reproductive medicine, particularly in major cities.¹² Of these institutions, 38 are located in Jiangsu province, ranking second in the country. Participants were recruited by simple random sampling from patients who attended fertility centres in three three-tier hospitals located in Jiangsu Province. The inclusion criteria for participants were: women aged 20–45 years, suffering from infertility, seeking medical assistance for infertility, having good communication skills, and voluntary participation in the questionnaire study. Patients who were interested and eligible to participate in the study will have a face-to-face questionnaire and receive a gift. The hospitals were Nanjing Drum Tower Hospital, Jiangsu Province Hospital and Changzhou No.2 People’s Hospital. The sample size was determined by Orme’s formula (n>500*c/(t*a)), which required a minimum of 84 respondents (with a maximum of three levels, nine choice sets and two alternatives).¹³ There were three versions of the questionnaire, requiring a minimum of 252 questionnaires. Considering the response rate and tolerance error, we aimed to investigate 450 patients to ensure reliable statistical estimates and subsequent heterogeneity analyses. The anonymous survey was carried out between December 2022 and February 2023. Prior to survey participation, respondents were fully informed of its purpose and contents by trained researchers. All respondents gave informed consent for the study.

Data analysis

This study uses DCE as a research method and is based on a random utility framework to elicit the stated preferences of patients. Cost attributes are treated as continuous variables, whereas other attributes are represented by dummy variables. In the sensitivity analysis, linear coding was applied for all but participation in treatment decision-making (TDM). The mixed logit model (MIXL) assumes that the coefficients on attribute levels follow a normal distribution to account for preference heterogeneity.¹⁴ The presence of significant positive or negative coefficients indicates a preference for certain attribute levels. An alternative-specific constant was included in the MIXL for unforced data. MIXL was used in Stata V.17 to analyse the passed consistency test data versus all valid data, as well as forced choice data versus unforced choice data.¹¹ The relative importance (RI) of attributes can be estimated by comparing the range of utilities for each attribute. Infertile patients’ WTP is indicated by the ratio of the coefficient of each non-cost attribute to that of the cost attribute. Additionally, the marginal rate of substitution (MRS) between non-cost attributes was assessed.¹⁵ We determined the number of categories based on log-likelihood, Akaike information criterion and Bayesian information criterion, and interpreted and named them.¹⁶ Subgroup analysis explored differences in preferences. Statistical significance was determined using p<0.05 (two-sided), and 95% CIs were calculated using the delta method. All statistical analyses were conducted using Stata, V.17.

Results

Respondent characteristics

In the course of the face-to-face investigation, 500 questionnaires were distributed to all participating women. Of these, 471 were returned, resulting in a response rate of 94.2% (471/500). However, six respondents did not complete the questionnaire in its entirety. Consequently, 465 questionnaires were available for analysis. The proportion of patients who passed the consistency test was 81.29%. The mean age of the 465 respondents was 31 years. As much as 88.0% of the respondents had a high school education level or above and 66.2% were employed. The proportion of average monthly household income between $742 to $2226 is higher than half. Regarding treatment experience, 91.2% expressed a positive desire to conceive. In addition, 88.4% of patients had been preparing for pregnancy for less than 5 years, and 55.1% had undergone relevant treatment previously. During the survey, 41.7% of patients had their first visit. 92.7% of the participants reported good health or above. Further details regarding the participants’ information are presented in online supplemental table S5.

Preference analysis

Between the complete sample and the sample that passed the consistency test, there was no significant difference in preference (see online supplemental table S6 for details). Table 2 presents the regression results based on MIXL for the data that passed the consistency test. All six attributes significantly influenced the respondents’ choices at least on one level (p<0.05). Specifically, the higher degree of participation in TDM, the higher live birth and clinical pregnancy rate, the lower risk of complications, the lower the cost were all associated with a greater likelihood of patients choosing a particular treatment option. This aligns with our initial hypothesis. Most estimated SDs were statistically significant (p<0.05), indicating the existence of preference heterogeneity.

Table 2

Regression results based on the mixed logit model

Attribute and level	Coefficient (SE)	SD (SE)	ORs	WTP ($)	95% CI
Cost (cont.*$1484)	−0.11** (0.03)	−0.34*** (0.04)	0.90
Participation in TDM (ref: no participation)
Partial participation	0.97*** (0.22)	0.77* (0.34)	2.64	13393.31**	4751.75	22 034.85
Full participation	1.19*** (0.22)	0.72** (0.26)	3.29	16442.73**	6719.66	26 165.82
Clinical pregnancy rate (ref: 20%)
40%	0.20 (0.27)	−0.74 (0.41)	1.22	2719.23	−5130.14	10 568.61
60%	1.07*** (0.15)	1.39*** (0.19)	2.91	14747.46**	5567.19	23 927.74
Live birth rate (ref: 15%)
35%	1.01** (0.36)	−2.36*** (0.54)	2.76	14026.32**	3593.04	24 459.60
55%	1.52*** (0.22)	1.05*** (0.26)	4.57	20996.82**	7616.20	34 377.45
Risk of neonatal complications (ref: 15%)
5%	0.52*** (0.11)	0.57** (0.18)	1.69	7245.10**	2323.28	12 166.92
10%	0.06 (0.19)	−0.13 (0.44)	1.06	786.26	−4586.54	6159.06
Risk of maternal complications (ref: 5%)
0.1%	0.43*** (0.12)	−0.99*** (0.18)	1.53	5891.14**	1553.39	10 228.89
2.5%	−0.07 (0.24)	−0.56 (0.57)	0.93	−984.81	−7495.67	5526.05
Constant	0.35* (0.14)	2.23*** (0.23)	1.42	–	–	–
Log likelihood	−1665.42
AIC	3378.84
BIC	3539.82

Notes: *p<0.05; **p<0.01; ***p<0.001.

cont, continuous variable

AIC, Akaike information criterion; BIC, Bayesian information criterion; TDM, treatment decision-making; WTP, willingness to pay.

Among the six attributes considered, treatment options with relatively high live birth rates have the highest preference weighting (coefficient=1.52), followed by the degree of participation in TDM (coefficient=1.19). The risk of maternal complications was of lesser importance among the six attributes noted. Increasing the live birth rate from 15% to 55% produces 3.5 times more utilities (1.52/0.43) than decreasing the risk of maternal complications from 5% to 0.1%. In terms of ORs, the odds of choosing treatments offering 35% and 55% increases in the live birth rate were 2.76 and 4.57 times the odds of choosing treatments offering the reference level live birth rate. As shown in online supplemental table S7, a comparison between the results of the two-step opt-out models and forced choice models revealed no significant differences. The exit option did not impact choice significantly.

WTP and MRS

The highest WTP was observed when increasing the live birth rates. Patients were willing to pay $20 996.82 in order to increase live birth rate from 15% to 55%. Additionally, they were willing to pay $16 442.73 if the participation in TDM increased from a lower level to an adequate level. Additionally, they were willing to pay relatively little to reduce the risk of complications.

The results of the linear risk specification model are provided in online supplemental table S8. Patients were willing to accept a 0.69% (95% CI 0.42% to 0.97%) increase in the risk of neonatal complications to raise the live birth rate by 1%. On the other hand, this implies that patients would agree to a reduction of 1.44% (95% CI 1.03% to 2.38%) in the live birth rate in exchange for a 1% decrease in the risk of neonatal complications.

Importance ranking

The RI of attributes is presented in figure 1. Live birth rate was the most important attribute (RI=29.05%), followed by participation in TDM (RI=22.75%), while out-of-pocket cost and risk of maternal complications were considered less important.

Figure 1. Relative importance of attributes. TDM, treatment decision-making.

Predicted uptake rates for different scenarios

Figure 2 presents the predicted uptake results when changing specific attribute levels. We set the cost of $7422, no participation in TDM, 20% clinical pregnancy rate, 15% live birth rate, 15% risk of neonatal complications and 5% risk of maternal complications as the baseline. Within one attribute, patient preference had the greatest impact due to an increase in live birth rate from 15% to 55%, leading to 64.08% increase in the probability of choosing this treatment option. Within multiple attributes, the most favourable treatment option is ‘③+④+⑤+⑥’, featuring a higher degree of participation in TDM, a higher live birth and clinical pregnancy rate, and a lower risk of neonatal complication, and the choosing probability can increase 97.32%.

Figure 2. Policy simulation analysis of treatment option with specific attributes. TDM, treatment decision-making.

Preference heterogeneity analysis

Table 3 presents a latent class model with two classes. In class 1, respondents showed a significant preference for non-cost attributes (p<0.05), placing greater emphasis on live birth and clinical pregnancy rates than on cost. Consequently, we have defined this class as ‘outcome driven’. The RI of each attribute was ranked in the same order as in the main effects analysis. The coefficient for opt-out was −0.18, indicating that respondents were less likely to abandon treatment in the future. In class 2, labelled ‘cost driven’, there was no significant difference in the preferences of respondents for non-cost factors (p>0.05). Respondents preferred treatment options that were inexpensive and involved full participation in TDM. Subsequent analysis was conducted to explore the factors that influence patient preferences. The results can be found in online supplemental table S9. With class 2 as a reference, patients who lived out of the town (outside of the city or province) and had an average monthly household income exceeding $742 were more susceptible to being categorised as class 1, called ‘outcome driven’.

Table 3

Regression results based on the latent class model

Attribute and level	Class 1 (outcome focused)		Class 2 (cost focused)
	Coefficient	SE	Coefficient	SE
Cost (cont.*$1484)	0.03	0.01	−0.36***	0.11
Participation in TDM (ref: no participation)
Partial participation	0.60***	0.12	0.20	0.69
Full participation	0.70***	0.12	0.83**	0.82
Clinical pregnancy rate (ref: 20%)
40%	0.29	0.15	0.28	1.03
60%	0.63***	0.07	0.45	0.39
Live birth rate (ref: 15%)
35%	0.54**	0.18	0.67	0.97
55%	0.86***	0.11	0.65	0.78
Risk of neonatal complications (ref: 15%)
5%	0.35***	0.07	0.30	0.38
10%	0.20	0.12	−0.53	0.96
Risk of maternal complications (ref: 5%)
0.1%	0.21**	0.06	0.05	0.36
2.5%	0.25	0.13	−0.40	0.96
Constant	−0.18**	0.05	1.62***	0.63
Class share	28.1%		71.9%
Log likelihood	−1740.91
AIC	3549.09
BIC	3716.77

Notes: *p<0.05; **p<0.01; ***p<0.001.

AIC, Akaike information criterion; BIC, Bayesian information criterion; cont, continuous variable; TDM, treatment decision-making.

The detailed findings of the subgroup analyses are presented in online supplemental table S10. The research found that patient preferences varied according to their age and monthly household income. Women under the age of 35 with infertility were more inclined towards treatments that were inexpensive and allowed greater participation in TDM, compared with women over 35 years of age. The analysis also showed that respondents with a monthly income of $2226 or more also significantly varied in these two attributes.

Discussion

This study investigated the preferences of infertile patients for ART in Jiangsu, China. The results from the DCE study indicate the significant preference for all six attributes. When seeking fertility treatment, patients prioritised live birth rates above all else, followed by the level of participation in TDM. Furthermore, the analysis of heterogeneity revealed that the preferences of infertile patients were linked to their age, income and geographical location.

The majority of research on the preferences of infertile patients has been conducted in Western countries, particularly Australia and the Netherlands.^{9 16 17} Patient preferences are remarkably consistent across countries and this overlap demonstrates the applicability of the study. To our knowledge, this research represents the first study employing a DCE to measure patients’ choice preferences and WTP concerning ART in China. While the results demonstrate considerable uniformity with international findings, some notable differences exist. A single study conducted in China using rating-based conjoint analysis to investigate the preferences of infertile patients revealed that patients placed the greatest value on their doctor’s attitude, placing it approximately 10% more important than the success rate.¹⁸ This discrepancy may be attributed to the evolution of the healthcare system, which has been undergoing significant changes in recent years. Consequently, patient experience and value-based healthcare must be prioritised, to enable ART to transition from a consumer healthcare role. It should be transformed into a core medical technology with a focus on patient experience. In a large-scale cross-national DCE covering the USA, the UK, China and others, the preference for cost was not significant among patients with low fertility or experience of ART in the Chinese sample.¹⁹ In contrast, in the USA, the cost attribute was almost as important as the live birth rate attribute. This may be attributed to the fact that health insurance reimbursement policies were not available in China at the time, resulting in significant out-of-pocket costs. With the high incidence of infertility and the significant economic burden, China’s infertility economic support and medical insurance policies are still in the early stages of development compared with other countries. This is crucial for discussing the findings in the current healthcare environment in China.

The demand for higher-quality healthcare among domestic patients is expanding, encompassing not only technological advances and clinical outcomes but also the entire treatment process. Our study showed that patients prefer to participate fully in TDM, ranking this attribute second in importance after live birth rates. Physician-patient shared decision-making is a collaborative model used to convey treatment information, values and preferences. This model is increasingly recognised as the optimal framework for involving both patients and physicians in clinical care.²⁰ In the context of diseases with uncertain treatment characteristics, there is a general trend towards personalised treatment plans. Decision-making regarding women’s reproductive health involves distinct interactions between physicians, women and partners.²¹ Furthermore, women, regardless of their health literacy, exhibit a considerable interest in acquiring knowledge related to maternity information and are highly motivated to do so independently.²² From a medical viewpoint, while some patient preferences may not be feasible, clinicians should attentively heed patients and impart appropriate clarifications to advance patient cooperation and gratification with therapy.^{5 23} Additionally, the pivotal role of the family in decision-making, influenced by traditional Chinese culture and other societal factors, is a distinctive feature of modern Chinese society. Thus, family centred decision-making is equally significant.²⁴ Integrating choice preferences into diagnostic and therapeutic decisions presents challenges for individualised personal decision assistants in practice.

This study was conducted in reproduction centres, which may limit the selection of the study population. The questionnaire was administered only to women who were in outpatient counselling or undergoing ART. This group of respondents may have a strong willingness to be treated and the financial capacity to afford such treatments. According to statistics, over 3 million ART procedures are performed globally each year, with an annual demand exceeding 10 million, yet only 46.5% of couples with infertility actively seek treatment.^{25 26} So, the study’s WTP results must be analysed critically. It is worth noting that respondents’ WTP ranges from 0.80 times to 2.83 times the per capita disposable income of the population in Jiangsu in 2022 ($7401). Moreover, studies have indicated that the estimated price elasticity of fertility treatments is very low, suggesting that women are willing to pay a high cost for fertility treatments regardless of their income level or ability to pay.¹⁶ In many countries, particularly middle-income developing nations, individuals usually bear the cost of fertility treatments, resulting in significant catastrophic health payments.²⁷ This impacts the equity of patients’ access to fertility care, with financial barriers being a major reason for patients’ inability to access treatment. In this study, 71.9% of the population holds a ‘cost driven’ preference, meaning that they prefer lower-cost treatments. Subgroup analyses indicated that women under the age of 35 years who experienced infertility preferred low-cost options, and this population segment is in their peak reproductive years.²⁸ In addition, the sunk costs of withdrawing from treatment, or the unquantifiable psychological burden,²⁹ are also components of the overall treatment cost. Significant increases in the number of transplant cycles and success rates have been observed in overseas studies following the introduction of government health insurance reimbursement coverage.³⁰ The government’s provision of appropriate fertility treatment subsidies to low-income infertile couples, or the inclusion of treatment in health insurance plans, has been demonstrated to be an effective measure to ensure that more patients with a desire to procreate have access to fair, accessible and affordable healthcare. This approach can also help ease the doctor-patient relationship and increase fertility rates. The low penetration of ART also suggests potential for future growth and market expansion within the population policy framework. Therefore, the government must implement inclusive policies to support fertility, sharing the cost of healthcare and contributing towards achieving a satisfactory level of fertility.

This study also has several limitations. First, this study does not consider couple infertility but only the woman’s answers, which may lead to an inadequate conclusion. Second, although we identified and selected attributes for this study mostly through literature review and empirical study, we cannot guarantee that all crucial attributes related to ART were included. Lastly, like with most DCE studies, it was not possible to confirm the external validity. However, we implemented consistency tests and a two-step choice process to validate the internal validity.

Conclusion

Understanding patient preferences is crucial for achieving patient-centred care. Patient preferences are primarily driven by live birth rates, pregnancy rates and the participation in TDM, which also influence WTP. Future studies need to examine the WTP and price elasticity of fertility treatments, for individuals who have not undergone ART, across multiple treatment cycles. Patient preferences should be taken into account when implementing any policy to ensure that treatments are aligned with patient needs and expectations.

The authors thank the respondents for taking part in the survey.

Data availability statement

Data are available upon reasonable request.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

This study involves human participants and was approved by the ethics committee of Nanjing Medical University (ethics approval number 2020(103)). Participants gave informed consent to participate in the study before taking part.

Footnote

Contributors JC, YB, LZ and XL conceptualised and designed the study. JC, YL, YS, XQ, KF, BW and HD collected and rechecked the data. JC and LZ conducted the data analysis. JC drafted the manuscript. LZ and XL provided critical revision of the manuscript. XL supervised the study and provided administrative support. XL is responsible for the overall content as guarantor.

Funding This work was supported by the National Natural Science Foundation of China (XL, grant number 72074123), (LZ, grant number 72304150) and the Jiangsu Province Postgraduate Research and Practice Innovation Program (JC, grant number KYCX23_1901).

Competing interests None declared.

Patient and public involvement Patients and/or the public were involved in the design, or conduct, or reporting, or dissemination plans of this research. Refer to the Methods section for further details.

Provenance and peer review Not commissioned; externally peer reviewed.

Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.

References

¹ Mascarenhas MN, Flaxman SR, Boerma T, et al. National, Regional, and Global Trends in Infertility Prevalence Since 1990: A Systematic Analysis of 277 Health Surveys. PLoS Med 2012; 9: e1001356. doi:10.1371/journal.pmed.1001356

² Qiao J, Wang Y, Li X, et al. A Lancet Commission on 70 years of women’s reproductive, maternal, newborn, child, and adolescent health in China. Lancet 2021; 397: 2497–536. doi:10.1016/S0140-6736(20)32708-2

³ Kushnir VA, Barad DH, Albertini DF, et al. Systematic review of worldwide trends in assisted reproductive technology 2004-2013. Reprod Biol Endocrinol 2017; 15: 6. doi:10.1186/s12958-016-0225-2

⁴ Tatum M. China’s fertility treatment boom. Lancet 2020; 396: 1622–3. doi:10.1016/S0140-6736(20)32475-2

⁵ Cornelisse S, Vos MS, Groenewoud H, et al. Women’s preferences concerning IVF treatment: a discrete choice experiment with particular focus on embryo transfer policy. Hum Reprod Open 2022; 2022: hoac030. doi:10.1093/hropen/hoac030

⁶ Botha W, Donnolley N, Shanahan M, et al. Assessment of the societal and individual preferences for fertility treatment in Australia: study protocol for stated preference discrete choice experiments. BMJ Open 2018; 8: e020509. doi:10.1136/bmjopen-2017-020509

⁷ Keller E, Botha W, Chambers GM. What Features of Fertility Treatment do Patients Value? Price Elasticity and Willingness-to-Pay Values from a Discrete Choice Experiment. Appl Health Econ Health Policy 2023; 21: 91–107. doi:10.1007/s40258-022-00764-7

⁸ Skedgel C, Ralphs E, Finn E, et al. How Do People with Experience of Infertility Value Different Aspects of Assistive Reproductive Therapy? Results from a Multi-Country Discrete Choice Experiment. Patient 2022; 15: 459–72. doi:10.1007/s40271-021-00563-7

⁹ Reed Johnson F, Lancsar E, Marshall D, et al. Constructing experimental designs for discrete-choice experiments: report of the ISPOR Conjoint Analysis Experimental Design Good Research Practices Task Force. Value Health 2013; 16: 3–13. doi:10.1016/j.jval.2012.08.2223

¹⁰ Lancsar E, Louviere J. Conducting discrete choice experiments to inform healthcare decision making: a user’s guide. Pharmacoeconomics 2008; 26: 661–77. doi:10.2165/00019053-200826080-00004

¹¹ Bridges JFP, Hauber AB, Marshall D, et al. Conjoint analysis applications in health--a checklist: a report of the ISPOR Good Research Practices for Conjoint Analysis Task Force. Value Health 2011; 14: 403–13. doi:10.1016/j.jval.2010.11.013

¹² Hauber AB, González JM, Groothuis-Oudshoorn CGM, et al. Statistical Methods for the Analysis of Discrete Choice Experiments: A Report of the ISPOR Conjoint Analysis Good Research Practices Task Force. Value Health 2016; 19: 300–15. doi:10.1016/j.jval.2016.04.004

¹³ Ride J, Goranitis I, Meng Y, et al. A Reporting Checklist for Discrete Choice Experiments in Health: The DIRECT Checklist. Pharmacoeconomics 2024; 42: 1161–75. doi:10.1007/s40273-024-01431-6

¹⁴ Botha W, Donnolley N, Shanahan M, et al. Societal preferences for fertility treatment in Australia: a stated preference discrete choice experiment. J Med Econ 2019; 22: 95–107. doi:10.1080/13696998.2018.1549055

¹⁵ Veldwijk J, Lambooij MS, de Bekker-Grob EW, et al. The effect of including an opt-out option in discrete choice experiments. PLoS One 2014; 9: e111805. doi:10.1371/journal.pone.0111805

¹⁶ Wang Y, Zhai P, Jiang S, et al. Blood Donors’ Preferences Toward Incentives for Donation in China. JAMA Netw Open 2023; 6: e2318320. doi:10.1001/jamanetworkopen.2023.18320

¹⁷ He X, Wang X, Shen J, et al. Cost-effectiveness of preimplantation genetic testing for aneuploidy for women with subfertility in China: an economic evaluation using evidence from the CESE-PGS trial. BMC Pregnancy Childbirth 2023; 23: 254. doi:10.1186/s12884-023-05563-z

¹⁸ Cai QF, Wan F, Dong XY, et al. Fertility clinicians and infertile patients in China have different preferences in fertility care. Hum Reprod 2014; 29: 712–9. doi:10.1093/humrep/deu023

¹⁹ de Bekker-Grob EW, Donkers B, Jonker MF, et al. Sample Size Requirements for Discrete-Choice Experiments in Healthcare: a Practical Guide. Patient 2015; 8: 373–84. doi:10.1007/s40271-015-0118-z

²⁰ Veldwijk J, DiSantostefano RL, Janssen E, et al. Maximum Acceptable Risk Estimation Based on a Discrete Choice Experiment and a Probabilistic Threshold Technique. Patient 2023; 16: 641–53. doi:10.1007/s40271-023-00643-w

²¹ Tucker Edmonds B. Shared decision-making and decision support: their role in obstetrics and gynecology. Curr Opin Obstet Gynecol 2014; 26: 523–30. doi:10.1097/GCO.0000000000000120

²² Geng L, Shi Z, Chai X-Y, et al. Patient and clinician perspectives on shared decision-making in infertility treatment: A qualitative study. Patient Educ Couns 2023; 116: S0738-3991(23)00328-2. doi:10.1016/j.pec.2023.107948

²³ Fransen MP, Hopman ME, Murugesu L, et al. Preconception counselling for low health literate women: an exploration of determinants in the Netherlands. Reprod Health 2018; 15: 192. doi:10.1186/s12978-018-0617-1

²⁴ Klitzman R. Impediments to communication and relationships between infertility care providers and patients. BMC Womens Health 2018; 18: 84. doi:10.1186/s12905-018-0572-6

²⁵ Copp T, Kvesic D, Lieberman D, et al. “Your hopes can run away with your realistic expectations”: a qualitative study of women and men’s decision-making when undergoing multiple cycles of IVF. Hum Reprod Open 2020; 2020: hoaa059. doi:10.1093/hropen/hoaa059

²⁶ Gao S, Corrigan PW, Qin S, et al. Comparing Chinese and European American mental health decision making. J Ment Health 2019; 28: 141–7. doi:10.1080/09638237.2017.1417543

²⁷ Adamson GD, Zegers-Hochschild F, Dyer S. Global fertility care with assisted reproductive technology. Fertil Steril 2023; 120: 473–82. doi:10.1016/j.fertnstert.2023.01.013

²⁸ Zhou Z, Zheng D, Wu H, et al. Epidemiology of infertility in China: a population-based study. BJOG 2018; 125: 432–41. doi:10.1111/1471-0528.14966

²⁹ Dyer SJ, Vinoos L, Ataguba JE. Poor recovery of households from out-of-pocket payment for assisted reproductive technology. Hum Reprod 2017; 32: 2431–6. doi:10.1093/humrep/dex315

³⁰ Vitagliano A, Paffoni A, Viganò P. Does maternal age affect assisted reproduction technology success rates after euploid embryo transfer? A systematic review and meta-analysis. Fertil Steril 2023; 120: 251–65. doi:10.1016/j.fertnstert.2023.02.036