Preferences for group psychological interventions on perinatal depression in China: a discrete choice experiment

Abstract

Background

Perinatal depression (PND) is a public health issue that causes a considerable disease burden on health systems and patients’ families. Group psychological interventions are intended to prevent PND. This study aimed to identify maternal preferences for group psychological interventions and provide evidence for intervention design and implementation in China.

Methods

A discrete choice experiment (DCE) was conducted in southern China from January to February 2023. Based on a literature review, expert consultation, and qualitative interviews, six attributes of group psychological interventions were selected for the DCE: cost, qualification of care provider, care method (online/offline), care receiver, institution for care delivery and starting time of care. A mixed-logit regression model was used to analyze preferences, willingness to pay, and potential uptake rate.

Results

The survey included 517 participants. They reported preferring free interventions provided by psychologists, available online and offline, with their spouses, and at the hospital where they received routine maternity care. Participants had no significant preferences for starting time of care. Inviting spouses to attend the intervention was the most valued attribute. The preferences of the different subgroups of respondents were diverse.

Conclusion

Considering pregnant women’s needs and preferences when designing group psychological interventions will help to increase their willingness to participate in the intervention for PND prevention. Future maternal group psychological interventions could invite spouses and other informal family caregivers to accompany them, reduce or waive fees for pregnant women through financial support, and provide online interventions for pregnant women in the third trimester and postpartum period.

Supplementary Information

The online version contains supplementary material available at 10.1186/s13690-025-01643-y.

Text box 1. Contributions to the literature
1. WHO proposed integrating perinatal mental health care into primary maternal and child care in low - and middle-income countries. Psychological intervention designed based on maternal preference could promote utilization of the integrated care.
2. There is limited evidence on maternal preferences for group psychological intervention on perinatal depression, especially in developing countries.
3. Pregnant women prefer free interventions provided by psychologists, available online and offline, with their spouses, and at the hospital where they received routine maternity care. Inviting spouses to participate could increase their willingness to receive intervention in primary health institutions.

Background

Perinatal depression (PND) is a common perinatal mental health problem, with an estimated global prevalence of 26.3%, with high prevalence rates for both prenatal depression (28.5%) and postnatal depression (27.6%) in 2023 [1]. In China, the prevalence of PND was 16.3% in 2020 [2], which is higher than the prevalence of 11.4% in developed countries [3] in 2017. Social, psychological, and biological factors are associated with PND, such as a history of depression, stressful life events, vulnerable personality, and poor social support [4]. In addition to significant emotional and social functioning and physical consequences on patients [5, 6], PND is associated with less exclusive breastfeeding, infant sleep problems, poor maternal–infant bonding and interaction, increased rates of childhood mental disorders, and increased risk of significant developmental delays, which affect children’s physical, emotional, and neurological development, as well as intimate relationships between couples [7]. Therefore, PND imposes a significant burden on families and society.

PND prevention and treatment mainly include drug, physical, and psychological interventions. According to the 2022 WHO “Guide for Integration of Perinatal Mental Health in Maternal and Child Health Services,” psychological interventions can be applied to prevent PND in the general population, and the treatment of mild and moderate PND with few side effects and high-efficiency [8–10]. Different types of psychotherapy for PND have been developed and practiced [11], including cognitive behavioral interventions, interpersonal interventions, and mindfulness therapy [12–15], which can be delivered in individual, group, telephone, and guided self-help formats [16]. Compared to individual psychological interventions, the advantages of group psychological interventions include learning by observing others (i.e., vicarious learning), knowing and being comforted by others sharing the same difficulties, and practicing constructive solutions for interpersonal problems in a safe environment [17]. WHO's “Thinking Healthy: A Manual for Psychosocial Management of Perinatal Depression” advocates that group psychological interventions should be first-line management for PND and recommends integrating PND management into maternal and child programs in primary care. The economic effectiveness of integrated care models for perinatal mental health, which involve screening for perinatal mental health problems and low-intensity treatments administered by midwives and health visitors in collaboration with primary mental health services, is supported by studies in high-income countries [18]. For low- and middle-income countries, task sharing, a model in which community health workers are trained and supervised by mental health specialists to deliver psychosocial interventions in primary and community care, can help tackle the resource gap for mental health professionals [19, 20].

In 2009, China implemented the Equalization of Basic Public Health Services Program, which includes community health centers (CHCs) to provide free health management during pregnancy, including five free prenatal health assessments and follow-up visits during all three trimesters, postnatal visits in the 7th and 42nd days, and prevention of mother-to-child transmission services. In 2020, China's Health Commission issued the “Work Plan for Depression Prevention and Treatment,” which proposes including PND screening in routine maternity care in both hospitals and CHCs. When detected by general practitioners in CHCs, pregnant women with PND are referred to hospitals for further assessment, diagnosis, and treatment by psychiatrists or psychologists. China has strengthened the training of general practitioners in CHCs to screen and prevent PND.

However, many pregnant women with PND remain undetected or untreated in China [21]. According to a meta-analysis, the detection rate of PND among pregnant women was 17.84% [22]. Although free screening has been introduced, the implementation still has a long way to go without strict regulations. Chen et al. found that only a few doctors appear to care about the emotional states of pregnant women [23]. After completing the psychological assessment scale, few doctors provided patients with feedback, follow-up, or referrals to hospitals [24]. Meanwhile, 38.9% of pregnant women felt uncomfortable when asked to undergo PND screening, and 20.7% said they would answer dishonestly, as they did not think screening was necessary [25]. Moreover, even when PND was detected, 3 out of 5 women chose not to receive treatment [25–27]. From the perspective of care providers, although China's investment in mental health care continues to increase, shortages in mental health funding and low-qualified practitioners for PND (especially in CHCs) persist, resulting in inadequate provision of and poor access to PND care [28, 29]. From patients' perspective, many obstacles hinder them from receiving psychological interventions for PND. A previous study showed that 37.8% of patients were unable to become aware of their psychological problems themselves. Lack of mental health knowledge and literacy seriously reduces psychological care-seeking behavior and utilization of related care [30]. Additionally, most psychological care provided by qualified psychologists in hospitals is not covered by health insurance [23], which may reduce the care-seeking behavior of some patients. In addition, the serious stigma of mental illness in society affects the early identification and care-seeking behavior of some patients [31]. Only 5.4% of women with PND sought professional help in China, which was much lower than in developed countries (13.6% ~ 58%) [32]. Therefore, with the increasing investment in PND by the government in institutions and providers, measures are necessary to remove obstacles from the perspective of patients. Developing psychological interventions based on the preferences of pregnant women is promising for meeting their demands well and raising the uptake rate of psychological interventions [33, 34]. However, little evidence is available regarding the psychological care preferences of pregnant women.

A discrete choice experiment (DCE) is a quantitative method of eliciting preferences [35]. It is increasingly applied to inform policymaking regarding medications, health technologies, and health services from the perspective of consumers [36–38]. This study aimed to investigate pregnant women’s preferences and willingness to pay (WTP) for group psychological care using DCE. The results may enlighten the design of group psychological intervention.

Methods

Study setting

The research was conducted at Guangdong Maternal and Child Health Hospital (Yuexiu District) from January to February 2023. It is one of three districts of the Guangdong Maternal and Child Health Hospital in Guangzhou City. It provides more than 85, 000 outpatients and 4,000 delivery services annually. The hospital provides over two million outpatient services and 15,000 delivery services annually in all three districts. The three districts combined have only one Mental Health Department in Panyu District, with two psychotherapists providing professional psychological screening and assessment, individual counseling, group counseling, and marital relationship counseling for pregnant women. In 2022, the Mental Health Department provided 456 outpatient services, accounting for 0.02% of the total hospital outpatient services.

DCE development

Selection of attributes and levels

Attributes and levels of the DCE were selected based on a literature review; group discussions with 10 pregnant women [39] and a key informant interview (n = 8) with four psychologists, two psychiatrists, and two general practitioners. The four psychologists are from universities, two psychiatrists are from hospitals, and two general practitioners are from community health centers. All of the professionals participated in an interview session at the same time. The literature review found several factors influencing the utilization of psychological care for mothers in China [4], including the starting time of care, frequency of care delivery, length of care, method of care delivery (online/offline), content of care, institution for care delivery, qualifications of care providers, care receivers, and cost of care. In interviews and focus group discussions, three influencing factors (frequency of care delivery, length of care, and content of care) were excluded. The rationale behind this decision is that these factors should be determined based on physicians’ expertise rather than on pregnant women’s preferences. The cost of care enabled the estimation of WTP to improve the other five attributes and related levels. The three levels of each attribute were determined based on literature review and group discussion. Table 1 lists the six final attributes and their corresponding significance levels.

Table 1.

Attributes and levels of group psychological intervention

Attributes	Description	Levels
1. Institution for care delivery	You wish to receive group psychological intervention in ________	1.Community health centers
		2. Hospital where routine pregnancy care is provided
		3. Other hospitals
2. Qualification of care provider	You wish to receive group psychological intervention provided by ________	1. Nurse
		2. Psychologist
		3. Psychiatrist
3. Care receiver	You wish to participate in the group psychological intervention by ________	1. Yourself
		2. Yourself and spouse
		3.Yourself and informal carer for the child (other than you)
4. Care method (online/offline)	You would like to participate in the group psychological intervention ________	1. Online
		2. Offline
		3. Offline (before childbirth) + online (after childbirth)
5. Starting time of care	You would like to start the group psychological intervention in your________ (maternity stage)	1. First trimester
		2. Second trimester
		3. Third trimester
6. Cost	If each group psychological intervention lasts for 1 h with 15–20 pregnant women, you are willing to pay _______ (RMB) to attend	1. ¥0 ($0)
		2. ¥15 ($2.22)
		3. ¥30 ($4.45)

Experimental design and choice task development

A full factorial design of six attributes (with three levels each) produced 3⁶ = 729 scenarios and 265,356((729 × 728)/2) choice sets. A fractional factorial experimental design was used to generate a more manageable level of 18 choice sets. Each choice set consisted of alternatives 1 and 2 for which the attribute levels varied randomly. The 18 choice sets were randomly divided into two versions to avoid overloading respondents [40]. The questionnaire contained parts. Part 1 consisted of an introduction to the study, informed consent, sociodemographic information, and questions related to pregnancy. Part 2 included one choice set example and 10 choice sets. Thereafter, we conducted pre-test interviews with eight pregnant women, who demonstrated that they felt comfortable with the simple and friendly language used, understood the choice sets, and could make a logical decision between the two alternatives. According to WHO’s recommendations about integrating PND management into primary health care, we made CHCs the reference level in the attribute “institution for care delivery.” The reference levels of the other attributes were selected based on the lowest preferences of the pregnant women in the pre-test interviews.

Validity

If respondents chose alternatives 1 or 2 for all choice sets, the quality of the response was considered poor, and the records of these respondents were eliminated. Additionally, to further examine the quality of the responses, we conducted a consistency test, duplicating the first choice set and inserting it as the fifth choice set in the questionnaire [41, 42]. If the respondents made different choices for the 1st and 5th choice sets, they failed the consistency test. No exit option was set to prevent respondents from repeatedly selecting the opt-out item and lowering the response quality.

Sampling and data collection

In stated-preference methods, sample size calculations are based on the rule of thumb [43, 44], and the recommended minimum sample size was 500. We targeted a minimum sample size of 500 participants for this study. Considering a response rate of 60%, we invited 834 pregnant women to participate in the offline survey who were recruited using convenience sampling. The recruitment criteria included pregnant women who (1) received pregnancy management at Guangdong Maternal and Child Health Hospital, (2) were willing to participate, (3) were able to provide written informed consent, and (4) were able to complete the questionnaire with the help of trained investigators. Ethical approval was obtained from the Medical Ethics Committee of Sun Yat-sen University (Reference No.56). Supplementary Material Table 1 shows an example of this choice set.

Data analysis

STATA 16 was used for the statistical analyses. Random utility theory assumes that the utility associated with a good or service comprises the utility of its characteristics. [45] The utility acquired by individual n from alternative j is expressed as follows:

$$U_{\mathit{nj}}=V_{\mathit{nj}}+{\epsilon }_{\mathit{nj}},j=1,\cdots ,J$$

where $U_{\mathit{nj}}$ consists of a systematic component $V_{\mathit{nj}}$ and a random component ${\epsilon }_{\mathit{nj}}$. The systematic component $V_{\mathit{nj}}$ is a function of the observed care attribute $X_{\mathit{nj}}$ and observed individual characteristics of n. The random component ${\epsilon }_{\mathit{nj}}$ is related to the unobserved attributes or preference variations

$$V_{\mathit{nj}}={\beta }^{\prime }{\times }_{\mathit{nj}}+{\beta }^{\prime \prime }{C}_{n,}j=1,...,J$$

where $C_n$ represents the individual's characteristics n;$\beta _{}^{_{}^{\prime }}$ and $\beta _{}^{_{}^{\prime }_{}^{\prime }}$ are the vectors of the coefficients to be estimated.

Regression and willingness to pay

Relative importance was estimated for the attributes and levels in the DCE using a mixed logit regression analysis based on goodness of fit. The estimated coefficients in the regression model provided information on the direction and significance of the effects of changing attribute levels. However, they could not provide the value required for the reference level. Therefore, this study calculates WTP for an additional cost to receive a higher level of an attribute. The WTP for an attribute or level $x$ can be estimated as follows:

$$WTP\left(x\right)=\frac{\partial U/\partial x}{\partial U/\partial opt}$$

where opt denotes costs.

In addition, we conducted subgroup analyses of all respondents based on different maternity stages, age, employment status, per capita monthly family income, classification of pregnancy risk, and whether mental health services were used after pregnancy to elicit the specific subgroup preferences.

Changes in uptake rate

To provide evidence for integrating group psychological interventions with primary care in primary health institutions, we conducted a simulation to explore the potential uptake of participation in CHCs as the attribute levels changed. Random utility theory assumes that participants choose the option that provides them with the highest overall utility. When individual n is asked to choose between alternatives i and j, the probability of choosing i is determined as follows:

$$P_{\mathit{ni}}=Prob\left(U_{\mathit{ni}},>,U_{\mathit{nj}}\right)=Prob\left(V_{\mathit{ni}},+,{\epsilon }_{\mathit{ni}},>,V_{\mathit{nj}},+,{\epsilon }_{\mathit{nj}}\right)=Prob(V_{\mathit{ni}}-V_{\mathit{nj}}>{\epsilon }_{\mathit{nj}}-{\epsilon }_{\mathit{ni}})\forall i\ne j$$

When the random component ${\epsilon }_{\mathit{nj}}$ is assumed an independent and identically distributed extreme value, the probability that individual n chooses alternative i can be estimated as follows:

$$P_{\mathit{ni}}=\frac{e^{v_{\mathit{ni}}}}{{\mathrm{\Sigma }}_j{e}^{V_{\mathit{nj}}}}=\frac{e_{}^{(\beta _{}^{_{}^{\prime }}_{x_{\mathit{ni}}}^{}+\beta _{}^{_{}^{\prime }_{}^{\prime }}{C}_n)}}{{\mathrm{\Sigma }}_{j}e_{}^{(\beta _{}^{_{}^{\prime }}_{x_{\mathit{nj}}}^{}+\beta _{}^{_{}^{\prime }_{}^{\prime }}{C}_n)}}$$

where $X$ is a vector of attribute coefficients. If the levels related to the alternative $i$ are altered, the probability of receiving care using the previously preferred alternative $k$ changes, as follows:

$$Impactofpolicychange=P_{\mathit{ni}}-P_{\mathit{nk}},i\ne j$$

Results

Respondent characteristics

Of the 834 invited participants, 517 provided informed consent and completed the survey, corresponding to a response rate of 62%. Of those, 95 failed the consistency test. In the group of 422 respondents who passed the consistency test, 12 chose “alternative 1” and 8 chose “alternative 2” for all choice tasks. Sensitivity analyses that excluded respondents who failed the consistency test or chose the same alternative did not alter the results. At the same time, the AIC and BIC values of the model, excluding 115 samples (95 + 12 + 8) showed a better fit than the original model.( see Supplementary Material Table 2) Hence, we excluded 115 respondents and performed further analyses with 402 respondents.

Table 2 shows the characteristics of participants. The mean age in the sample was 31.26 years. A large proportion (73%) of the respondents were in their second and third trimesters. Two-thirds of the respondents had a college degree or higher, and more than half had jobs (62%) with a monthly family income of less than ¥100,00 ($1482.32, at an exchange rate of 1USD to 6.7462 CNY on January 1, 2023, when the survey started) per person (62%). Nearly half of the participants had no or low risk of pregnancy. Overall, 90% of the respondents reported they had not received mental health services during their current pregnancy.

Table 2.

Sociodemographic characteristics of the respondents

Characteristic	Respondents		Non-Respondents		X2	P
	(n = 402)		(n = 115)
	n	Percentage(%)	n	Percentage(%)
Age(year), mean ± SD	31.26 ± 4.36		31.56 ± 4.28
Age (year)
≤ 25	34	8	9	8	0.055	0.973
26–35	300	75	86	75
> 35	68	17	20	17
Maternity stage
First trimester	48	12	12	10	1.768	0.622
Second trimester	147	37	38	33
Third trimester	146	36	42	37
Postpartum	61	15	23	20
Marital status
Married	386	96	113	98	1.336	0.277
Unmarried	16	4	2	2
Education level
High middle school or below	137	34	47	41	1.798	0.180
College degree or above	265	66	68	59
Employment status
Employed	251	62	79	69	1.517	0.218
Unemployed	151	38	36	31
Monthly family income (Per capita)
< ¥5000 ($741.16)	77	19	26	23	6.178	0.046
¥5001–10000 ($741.31–1482.32)	174	43	35	30
> ¥10,001($1482.46)	151	38	54	47
Pregnancy risk level
No or low risk	170	42	55	48	1.116	0.291
Middle or high risk	232	58	60	52
Whether mental health services were received during this pregnancy
No	363	90	105	91	0.105	0.745
Yes	39	10	10	9

Preference and willingness to pay for group psychological intervention for PND

Table 3 presents the regression results and monetary valuations for each attribute and level. Aside from “Starting time of care,” the coefficients of the other attributes were statistically significant. Participants preferred free intervention provided by psychologists (β = 0.384, p < 0.001) online and offline (β = 0.297, p < 0.001), with their spouses (β = 0.584, p < 0.001) in maternal and child health hospitals where they are receiving routine maternity care (β = 0.209, p < 0.01). The estimated WTP for changes in attributes was the highest for care receivers. Compared to receiving interventions by themselves, participants were willing to pay an extra ¥23.83 ($3.53) and ¥10.44 ($1.55) for care if their spouses and family members, respectively, were also allowed to attend the intervention. Additionally, participants reported the following monetary values regarding WTP for changes to certain attributes and levels: care provided by psychologists instead of nurses (¥15.65/$2.32), online instead of offline care after childbirth (¥12.12/$1.8), care provided by a psychiatrist instead of nurses (¥10.71/$1.59), receive care in hospitals where routine maternity care is conducted instead of CHCs (¥8.52/$1.26), and online care both before and after childbirth (¥7.07/$1.05).

Table 3.

Regression results and WTP^a for group psychological intervention preventing PND

Attributes and levels	Coefficient means	Robust S.E	WTP (¥)	95%CI (¥)
Constant	−0.04	0.07	/	/
Institution for care delivery: Community health centers (refb)
Hospital where routine pregnancy care is provided	0.21**	0.06	8.52($1.26)	2.96($0.44)	14.08($2.09)
Other hospitals	−0.35**	0.07	−14.20(-$2.10)	−19.54(-$2.90)	−8.86(-$1.31)
Qualification of care providers: Nurse (refb)
Psychologist	0.38**	0.06	15.65($2.32)	10.44($1.55)	20.87($3.09)
Psychiatrist	0.26**	0.06	10.71($1.59)	5.82($0.86)	15.60($2.31)
Care receiver: Yourself (refb)
Yourself and spouse	0.58**	0.08	23.83($3.53)	16.70($2.48)	30.96($4.59)
Yourself and informal carer for the child (other than you)	0.26**	0.07	10.44($1.55)	5.16($0.76)	15.73($2.33)
Care method (online/offline): Offline (refb)
Online	0.17**	0.06	7.07($1.05)	2.09($0.31)	12.04($1.78)
Offline (before childbirth) + online (after childbirth)	0.30**	0.06	12.12($1.80)	7.04($1.04)	17.20($2.55)
Starting time of care: Third trimester (refb)
First trimester	−0.05	0.07	−2.02(-$0.30)	−7.89(-$1.17)	3.85($0.57)
Second trimester	−0.08	0.08	−3.18(-$0.47)	−9.55(-$1.42)	3.19($0.47)
Cost	−0.02	0.00	-	−0.03(-$0.00)	−0.02(-$0.00)

^aWTP is short for willingness to pay

^bref, reference that reflects the reference level for each attribute

^**p < 0.01

Subgroup analysis of preferences and willingness to pay for group psychological intervention for PND

Tables 4, 5 and 6 show the results of the analyses of preferences and WTP for group psychological interventions of pregnant women with differences in maternity stage, age group, and risk of pregnancy. Compared to pregnant women in other stages, pregnant women in the third trimester were more likely to want to participate in interventions at hospitals where they received routine maternity care. Women who were in the third trimester or postpartum were more willing to participate in online than offline interventions after childbirth. Most of the pregnant women aged 26–35 (74.6%) were more willing to receive online interventions before and after childbirth or online after childbirth and offline before childbirth than only offline care. Pregnant women with middle or high risks of pregnancy prefer interventions in hospitals to CHCs. Pregnant women who had received mental health services during their current pregnancy were more willing to receive online interventions provided by hospitals where they received routine maternity care than offline interventions provided by CHCs.

Table 4.

Regression results and WTP^a for group psychological intervention preventing PND (subgroup analysis: maternity stage)

	Maternity stage
Attributes	First trimester (n = 48)		Second Trimester (n = 147)		Third trimester (n = 146)		Postpartum (n = 61)
	Coefficient	WTP	Coefficient	WTP	Coefficient	WTP	Coefficient	WTP
Constant	−0.06		−0.09		0.02		−0.10
Institution for care delivery: Community health centers (refb)
Hospital where routine pregnancy care is provided	0.16	7.38($1.09)	0.21	6.11($0.91)	0.28*	11.02($1.63)	0.20	14.83($2.20)
Other hospitals	−0.46*	−21.96(-$3.26)	−0.27*	−8.13(-$1.21)	−0.42**	−16.55(-$2.45)	−0.34	−25.29(-$3.75)
Qualification of care providers: Nurse (refb)
Psychologist	0.16	7.80($1.16)	0.39**	11.50($1.70)	0.46**	18.04($2.67)	0.51**	37.70($5.59)
Psychiatrist	−0.03	−1.39(-$0.21)	0.20*	5.88($0.87)	0.39**	15.36($2.28)	0.45**	33.36($4.95)
Care receiver: Yourself (refb)
Yourself and spouse	0.36	17.14($2.54)	0.72**	21.40($3.17)	0.66**	26.12($3.87)	0.49*	36.51($5.41)
Yourself and informal carer for the child (other than you)	0.32	15.26($2.26)	0.29*	8.56($1.27)	0.25*	10.09($1.50)	0.23	16.90($2.51)
Care method (online/offline): Offline (refb)
Online	0.14	6.55($0.97)	0.04	1.13($0.17)	0.24*	9.46($1.40)	0.31	22.61($3.35)
Offline (before childbirth) + online (after childbirth)	0.30	14.37($2.13)	0.13	3.78($0.56)	0.39**	15.63($2.32)	0.44**	32.49($4.82)
Starting time of care: Third trimester (refb)
First trimester	−1.85	−8.79(-$1.30)	0.00	−0.06(-$0.01)	−0.16	−6.39(-$0.95)	0.15	11.38($1.69)
Second trimester	0.03	1.62($0.24)	−0.04	−1.03(-$0.15)	−0.25	−10.03(-$1.49)	0.08	5.64($0.84)
Cost	−0.02**	-	−0.03**	-	−0.03**	-	−0.01	-

^aWTP is short for willingness to pay

^bref, reference that reflects the reference level for each attribute

^*p < 0.05, **p < 0.01

Table 5.

Regression results and WTP^a for group psychological intervention preventing PND (subgroup analysis: age)

	Age (years)
Attributes	≤ 25 (n = 34)		26–35 (n = 300)		> 35(n = 68)
	Coefficient	WTP	Coefficient	WTP	Coefficient	WTP
Constant	0.30		−0.10		−0.05
Institution for care delivery: Community health centers (refb)
Hospital where routine pregnancy care is provided	−0.04	−1.58(-$0.23)	0.27**	11.07($1.64)	0.07	2.05($0.30)
Other hospitals	−0.25	−11.15(-$1.65)	−0.36**	−14.77(-$2.19)	−0.46*	−14.00(-$2.08)
Qualification of care providers: Nurse (refb)
Psychologist	0.37	16.44($2.44)	0.39**	15.97($2.37)	0.53**	16.25($2.41)
Psychiatrist	0.33	14.61($2.17)	0.25**	10.38($1.54)	0.34*	10.47($1.55)
Care receiver: Yourself (refb)
Yourself and spouse	1.08**	48.03($7.12)	0.54**	22.05($3.27)	0.65**	19.88($2.95)
Yourself and informal carer for the child (other than you)	0.47	20.89($3.10)	0.20**	8.27($1.23)	0.41*	12.55($1.86)
Care method (online/offline): Offline (refb)
Online	0.10	4.43($0.66)	0.18*	7.33($1.09)	0.17	5.32($0.79)
Offline (before childbirth) + online (after childbirth)	0.33	14.83($2.20)	0.31**	12.80($1.90)	0.25	7.78($1.15)
Starting time of care: Third trimester (refb)
First trimester	−0.06	−2.57(-$0.38)	−0.10	−4.28(-$0.63)	0.16	4.97($0.74)
Second trimester	0.01	0.43($0.06)	−0.14	−5.75(-$0.85)	0.03	0.89($0.13)
Cost	−0.02**	-	−0.02**	-	−0.03**	-

^aWTP is short for willingness to pay

^bref, reference that reflects the reference level for each attribute

^*p < 0.05, **p < 0.01

Table 6.

Regression results and WTP^a for group psychological intervention preventing PND (Subgroup analysis: Pregnancy risk level and whether they have received mental health services)

	Pregnancy risk level				Whether they have received mental health services
Attributes	Green (low risk)(n = 170)		Else (n = 232)		Yes (n = 39)		No (n = 363)
	Coefficient	WTP	Coefficient	WTP	Coefficient	WTP	Coefficient	WTP
Constant	−0.03		−0.05		−0.10		−0.04
Institution for care delivery: Community health centers (refb)
Hospital where routine pregnancy care is provided	0.19	7.08($1.05)	0.22*	9.27($1.37)	−0.03	−0.75(-$0.11)	0.23**	9.57($1.42)
Other hospitals	−0.35**	−12.60(-$1.87)	−0.38**	−15.88(-$2.35)	−0.34	−10.37(-$1.54)	−0.36**	−14.73(-$2.18)
Qualification of care providers: Nurse (refb)
Psychologist	0.34**	12.25($1.82)	0.43**	17.97($2.66)	0.57*	17.02($2.52)	0.37**	15.42($2.29)
Psychiatrist	0.11	3.89($0.58)	0.38**	16.14($2.39)	0.60**	18.20($2.70)	0.24**	9.78($1.45)
Care receiver: Yourself (refb)
Yourself and spouse	0.58**	21.04($3.12)	0.62**	26.11($3.87)	1.00**	30.18($4.47)	0.56**	23.03($3.41)
Yourself and informal carer for the child (other than you)	0.35**	12.76($1.89)	0.21*	8.73($1.29)	0.41	12.34($1.83)	0.25**	10.31($1.53)
Care method (online/offline): Offline (refb)
Online	0.07	2.64($0.39)	0.25**	10.59($1.57)	0.45	13.45($1.99)	0.16*	6.38($0.95)
Offline (before childbirth) + online (after childbirth)	0.25**	9.09($1.35)	0.34**	14.48($2.15)	0.40	11.90($1.76)	0.29**	12.12($1.80)
Starting time of care: Third trimester (refb)
First trimester	−0.02	−0.78(-$0.12)	−0.06	−2.71(-$0.40)	−0.14	−4.21(-$0.62)	−0.04	−1.75(-$0.26)
Second trimester	−0.02	−0.79(-$0.12)	−0.12	−4.88(-$0.72)	0.01	0.21($0.03)	−0.08	−3.45(-$0.51)
Cost	−0.03**	-	−0.02**	-	−0.03**	-	−0.02**	-

^aWTP is short for willingness to pay

^bref, reference that reflects the reference level for each attribute

^*p < 0.05,**p < 0.01

Potential uptake rate

According to recommendations from WHO and policies in China, integrating the management of PND into maternal and child health programs in primary care is cost-effective. Therefore, we aimed to predict the uptake rate of interventions in CHCs and explored scenarios to increase it. As shown in Fig. 1, the probability of participants receiving a baseline intervention provided by nurses at CHCs offline at a cost of ¥30 ($4.45) and attended by only the participants themselves was 44.8%. The best strategies for increasing the potential uptake rate of interventions at CHCs were reducing the cost to ¥0($0) (62.86%), followed by allowing spouses to also attend interventions (59.27%), providing interventions facilitated by psychologists (54.36%), reducing the cost to ¥15($2.22) (53.96%), and providing interventions offline before childbirth and online after childbirth (52.2%).

Fig. 1.

Probabilities of participating in group psychological intervention at community health centers

Discussion

This study measures the importance pregnant women assign to each attribute of the group psychological intervention for preventing PND and their WTP, and explores how choice probabilities of CHCs vary in changes in attribute levels. In order to assess the relative importance of different attributes and levels of group psychological intervention for pregnant women, this study investigated the attribute level preference, willingness to pay and uptake rate of pregnant women for group psychological intervention.

Main findings and comparison with other studies

This study’s findings suggest that pregnant women prefer to attend interventions with their spouses or informal caregivers. Previous studies have shown that spousal synchronized empowering education is effective in reducing maternal depression [46]. The engagement of a spouse and other family members can strengthen pregnant women’s social support, which is a protective factor against PND [47–49]. However, in many cases, spouses are not formally invited to participate in psychological care provided by the hospitals [50, 51]. Pregnant women were more likely to accept group psychological interventions provided by psychologists or psychiatrists rather than by nurses. In most previous studies conducted in China, group psychological care or education was provided by nurses or midwives in primary healthcare institutions and hospitals [52, 53]. Psychologists and psychiatrists tend to provide individual psychological counseling and treatment. Some respondents reported concerns regarding nurses’ qualifications for facilitating group psychological interventions. In addition, considering convenience and effectiveness, most pregnant women reported preferring offline interventions before childbirth and online interventions after childbirth. Currently, most health education and related group psychological care programs are conducted offline before childbirth [54, 55].

We found that respondents preferred group psychological interventions held in the hospitals where they received routine maternity care. On the one hand, receiving the intervention in hospitals for routine maternity care is more convenient than in other hospitals. On the other hand, hospitals for routine maternity care generally are more capable for mental health care than CHCs. This finding shows that the preferred institution is not the one recommended by WHO or national policy. As reported in previous studies, psychological care for pregnant women was predominantly provided in hospitals in China, rather than CHCs [56–59]. Despite the policies recommending screening for PND in CHCs, a qualitative study revealed that less than 30% of CHC care providers in CHCs had received training on screening and referral for PND [60].CHC providers’ qualifications and the lack of trust pregnant women have for them might lead to the low utilization and provision of related psychological screening, education, and intervention in CHCs in China [61, 62]. With varied levels of all attributes, the highest uptake rate for group psychological interventions in CHCs was 62.86%. Providing free interventions was the best measure, with the highest uptake rate. Inviting spouses to accompany pregnant women may be the most cost-effective measure, increasing the uptake rate to 59.27%. However, the proportion of spouses participating in antenatal care for pregnant women is low, with only 9.22% participating in all antenatal care activities for pregnant women. The reasons for spouses not participating in antenatal care activities include a lack of time (28.09%), feeling that it was the pregnant woman's business (25.84%), the pregnant woman did not ask them to (15.73%), having someone else to take care of the pregnant woman (6.74%), and being separated before childbirth (1.12%). [63]. Although providing interventions by psychologists instead of nurses is the third most effective measure, it is difficult for CHCs to implement. As reported in a study on mental health resources in China, by the end of 2015, there were 122,309 mental health professionals in China, but only 3,738 in primary healthcare institutions, accounting for 3.06% [64]. With the development of information technology, most CHCs can now provide online interventions through electronic information systems or official social media platforms. At present, most pregnant women’s courses are conducted through live broadcast software and hospital WeChat public accounts. Research has shown that the participation rate, correct answer rate, and teaching satisfaction scores of live broadcast courses are higher than those of offline courses [65, 66].

Subgroup analysis showed that, compared to pregnant women in the other trimesters, those in the third trimester were more likely to receive group psychological intervention in the hospital where they received routine maternity care. Pregnant women are expected to receive more routine prenatal care in the third trimester than in the first and second and face more inconvenience when traveling to other institutions. Therefore, they prioritize convenience and safety. Pregnant women in the 26–35 age group, accounting for 74.6% of the total participants, were more willing to receive online care after childbirth. This is consistent with research showing that most younger people are willing to consider virtual services for reasons of flexibility, convenience, privacy, and anonymity, as well as social distancing and safety during the COVID-19 pandemic [67]. Participants would have previously received mental health care preferred hospitals to CHCs because existing mental health care is mainly provided by hospitals. In a previous study, Magaard reported that availability of care and having a regular medical doctor are associated with a higher likelihood of seeking mental health treatment (70).

Policy implications

The principal findings of this study provide evidence for designing and optimizing group psychological interventions for PND from the perspective of pregnant women and for promoting the integration of routine mental health care for them. First, invite spouses and other informal caregivers in the family to participate in group psychological interventions. This accompaniment, based on understanding and support, has the potential to motivate pregnant women to participate in interventions and take care of their mental health. Providing a small gift for the father-to-be or conducting interventions on weekends could motivate other family members to participate. Second, the capacity to provide mental health care in CHCs should be strengthened. Providing routine assessment and screening for PND during pregnancy is the basis for increasing awareness of PND in pregnant women and building trust with them. Furthermore, essential training in psychological counseling for some care providers in CHCs could promote homogenization of mental health care between hospitals and CHCs and serve more pregnant women. Third, adequate financial support from the government is necessary to strengthen care capacity and provide free interventions in CHCs. As CHCs are non-profit institutions in China, salaries and operating costs depend on the government. Dedicated funding for mental healthcare could motivate institutions and care providers to provide related care. Finally, a combination of online and offline interventions should be adopted, as this can not only meet the needs of the majority age group (26–35 years old) but also fulfill the expectations of women in their third trimester of pregnancy and postpartum.

Strengths and limitations

To our knowledge, this is the first stated preference study to explore maternal preferences for mental health care (group psychological intervention) in China, with further analysis of different subgroups. The study design (i.e., DCE), quantifies the relative importance of various attributes at different levels. The results provided evidence on how to develop and implement group psychological interventions in primary care settings in China and other countries with relatively weak primary and mental care for pregnant women. However, this study has three limitations. First, DCE may not include other important attributes. To make the DCE manageable, we included six attributes based on a literature review, key informant interviews with care providers, and group discussions with pregnant women. Further research could adopt other important attributes such as intervention content. Second, this study was conducted at only one hospital. Meanwhile, the sample recruited to answer the survey was selected by convenience, which might lead to limited external validity of this study. Therefore, caution should be exercised when this study's results and policy implications are extended to other health settings. For example, two-thirds of the respondents had a college degree or higher, which may not be representative of the population at large. These selection biases may affect the results and generalization of the results, as more educated women could be more aware of mental health problems and have a higher WTP. All of the above can lead to selection bias. Subsequent studies should conduct multicenter trials and survey wider population to ensure the representativeness of the samples and results. Third, caution should be exercised when extending the results and policy implications of this study to other cities or countries. The results reflect a combination of factors including health system background, health resources, and participant demands. However, the DCE design and the adopted attributes are applicable to other healthcare settings. Finally, similar to other DCE studies, we investigated respondents’ stated preferences when the intervention was not provided. Further research could compare the results with revealed preferences based on participants’ actual behavior.

Conclusion

Psychological interventions show promise for preventing PND in the general pregnant population, especially when it is integrated with routine maternity care in primary care settings. Pregnant women’s needs and preferences are the key features of people-centered integrated maternal care. This study elicited maternal preferences for group psychological care, calculated WTP for the preferred attributes and levels of care, and predicted the likelihood that pregnant women would receive care in CHCs. Based on these findings, inviting spouses and other informal caregivers in the family to participate in the group psychological intervention, strengthening the capacity for providing intervention in CHCs, adequate financial support from the government for strengthening care capacity and providing free interventions, and implementing online interventions for women in the third trimester and postpartum are crucial strategies for increasing the uptake rate for group psychological care in primary care settings and CHCs in China. This study contributes to the development and implementation of mental healthcare in primary care settings and provides a methodology for investigating maternal preferences for mental healthcare in other healthcare systems.

Abbreviations

PND

Perinatal depression

DCE

Discrete choice experiment

CHCs

Community health centers

WTP

Willingness to pay

Authors’ contributions

Methodology, Software, Validation, Formal analysis, Investigation, Data curation, Writing-original draft, Writing – review & editing, SW; Investigation, Data curation, Writing – original draft, Writing – review & editing, YS; Methodology, Writing – original draft, Writing – review & editing, Supervision, JX; Investigation, Resources, Writing – original draft, Writing – review & editing, Supervision, Project administration, YM; Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing-original draft, Writing – review & editing, Visualization, Supervision, Project administration, Funding acquisition, XW.

Funding

This project was funded by the China Medical Board(CMB),under grant no. (CMB21-437).

Data availability

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author.

Declarations

Ethics approval and consent to participate

The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethics Committee, School of Public Health, Sun Yat-sen University (Reference No.56). Informed consent was obtained from all individual participants included in the study.

Consent for publication

The manuscript doesn’t contain any individual person’s data in any form.

Competing interests

The authors declare no competing interests.