Positive emotion interventions as a non-pharmacological approach to boost colostrum secretion: a randomized controlled trial

Abstract

Pregnancy emotional states can influence postpartum colostrum delivery, and lactational capacity establishes 1–5 days postpartum. Breast milk production capacity is key to promoting breastfeeding. This study applied positive psychology interventions to late-pregnancy primiparous women, aiming to explore how emotional regulation affects postpartum lactation-related hormone changes and in-hospital breastfeeding outcomes. Participants were recruited from a gynecology and obstetrics hospital in China. Positive writing interventions were delivered via WeChat electronic questionnaires: the control group recorded “first three things” and the intervention group “three good things,” twice weekly. Key outcomes (lactation hormones, neonatal milk consumption, breastfeeding outcomes) were collected, and generalized estimating equations (GEE) were used to analyze repeated-measurement data. Jan–Aug 2022, 204 (95 intervention, 109 control) were finally analyzed after excluding dropouts. GEE analysis showed the intervention group had a significantly higher resilience score than the control group (15.55 vs. 15.01; Wald χ² = 4.53, P = 0.033). No significant differences were found between groups in anxiety, depression, positive affect and negative affect schedule (PANAS) scores, or breastfeeding scores. Post-delivery, prolactin (PRL) levels differed significantly between groups (P = 0.031), but insulin-like growth factor-1 (IGF-1) levels did not (P > 0.05). Breastfeeding behavior effects also showed no inter-group differences. This study shows the positive psychological intervention based on the “three good things” approach can effectively increase primiparous women’s psychological resilience and promote prolactin secretion, but has no significant effect on breastfeeding behaviors. Additionally, the intervention tends to improve their postnatal positive emotions and reduce anxiety and negative emotions. This study provides clinical evidence for promoting colostrum secretion via prenatal positive psychology interventions.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-46318-y.

Introduction

Breast milk, a natural nutritional substance produced by lactating women, is abundant in macronutrients and bioactive constituents, exerting a pivotal role in infant growth and immune development¹. In 2022, a joint report released by the World Health Organization (WHO) and the United Nations Children’s Fund (UNICEF) indicated that merely 44% of infants under six months of age were exclusively breastfed worldwide². Over the past two decades, global breastfeeding rates have seen almost no growth, while the use of formula has nearly doubled during the same period³. It is estimated that if breastfeeding were to reach a universal level, approximately 820,000 children’s lives could be saved annually⁴. Breastfeeding reduces child mortality primarily by providing bioactive components (e.g., antibodies, antimicrobial molecules) in breast milk to protect against life-threatening infections (e.g., pneumonia, diarrhea), the leading causes of child deaths^5–7. Additionally, exclusive breastfeeding avoids pathogen ingestion via contaminated foods/fluids, early initiation lowers neonatal mortality risk, and adequate nutrition from breastfeeding reduces undernutrition-related mortality^4,8,9. An increase of just 10% in the breastfeeding rate could translate into significant cost savings in children’s healthcare: $312 million in the United States, $7.8 million in the United Kingdom, $30 million in urban areas of China, and $1.8 million in Brazil¹⁰. According to the China Development Research Foundation, the exclusive breastfeeding rate in China for infants under six months was 29.2% in 2019¹¹. This rate is still far from the targets set by the “Chinese Children Development Compendium (2011–2020)” and the "National Nutrition Plan (2017–2030)," which aim for a 50% rate, as well as the global average of 43% and 37% in low- and middle-income countries¹¹. In conclusion, breastfeeding promotion and practice in China still have significant room for improvement.

How to take effective intervention measures to further improve breastfeeding is a question worthy of our key consideration. Common intervention methods include breastfeeding health education, economic incentives, peer support, early continuous contact, overcoming social barriers^12–14. To date, psychological theories and interventions have not successfully accounted for breastfeeding behaviors¹⁵. The research perspective is overly simplistic, as it primarily focuses on psychological and behavioral dimensions while failing to consider physiological aspects. Therefore, we aim to explore new approaches to explain and promote breastfeeding.

Lactation ability is one of the key factors of breastfeeding behavior. The secretion of breast milk during the initial 1–5 days postpartum plays a formative role in subsequent lactation capacity, while the volume of breast milk directly influences breastfeeding-related behaviors and outcomes¹⁶. Milk secretion is often regulated by levels of lactation-related hormones. Prolactin (PRL) is a protein hormone that is regulated by the neuroendocrine and secreted by the pituitary gland, which promotes milk secretion by the mammary glands¹⁷. Insulin-like growth factor 1 (IGF-1) is a peptide hormone that also promote milk secretion, primarily by inhibiting mammary gland serotonin synthesis¹⁸. In addition, breastfeeding behavior can lead to higher PRL level, with lactation-induced prolactin release associated with baseline hormone levels¹⁹. Therefore, interventions targeting lactation-related hormone secretion and breastfeeding behavior in the early postpartum period are important.

Previous studies have shown that emotional changes can affect PRL and IGF-1 levels. Maternal mental health during pregnancy plays a crucial role in subsequent lactation ability during the early postpartum period¹⁶. Significantly delayed and reduced postpartum lactation is often observed in individuals with mental health problems²⁰. The emergence of negative maternal emotions has been associated with decreased PRL levels, reduced lactation, and lower breastfeeding rates²¹. Prenatal anxiety has also been linked to declines in maternal IGF-1 levels²². Therefore, it is valuable to monitor and address the emotions of pregnant women in late pregnancy to promote lactogen production and successful breastfeeding.

Current studies on promoting breastfeeding mainly focus on behavioral interventions, with a lack of evidence on the effect of promoting lactation. Anxiety symptoms during pregnancy are negatively correlated with exclusive breastfeeding behavior and its persistence²³. Addressing anxiety symptoms during pregnancy may help improve breastfeeding rates²³. There is evidence that maternal psychological distress can weaken lactation ability and the effectiveness of breastfeeding behavior²⁴. Therefore, this study aims to explore the promotion of early postpartum lactation as well as exclusive breastfeeding behavior through positive psychology interventions.

Methods

Study design and setting

This study was a randomized controlled trial conducted in China, which has been registered on 07/10/2020 in the Chinese Clinical Trial Registry (Registration No.: ChiCTR2000038849). A full version of the trial protocol and statistical analysis plan is available at https://doi.org/10.1136/bmjopen-2022–066601. The pilot study was conducted at Wuyi First People’s Hospital in Zhejiang Province from October 2020 to May 2021. The main study was conducted at a Grade-A tertiary hospital affiliated with Zhejiang University, from January 2022 to August 2022. This hospital is designated as a baby-friendly facility, where strict prohibitions are in place to prevent expectant mothers from bringing bottles or formula into the department. The recruited participants were stratified according to the degree of family support for breastfeeding in the enrollment questionnaire, and then randomly divided into the intervention group and the control group, and the follow-up intervention measures were single-blind. Field personnel are responsible for recruiting participants, and experimental researchers are responsible for generating random numbers and hiding them in groups.

Participants

All participants voluntarily enrolled and signed an informed consent form. Primigravidae were recruited from an obstetrics clinic of one hospital affiliated with Zhejiang University. The inclusion criteria were defined as follows: singleton primipara; aged over 18 years; gestational age between 34 and 38 weeks; intact cognitive and comprehension functions; adept at using WeChat; free of any contraindications to breastfeeding; intending to deliver in the hospital; and having signed the informed consent form. Exclusion applies to pregnant women with severe illnesses or infectious diseases that render breastfeeding inappropriate, to those with cognitive impairments or mental disorders, and to fetuses diagnosed with congenital anomalies.

The sample size estimation was calculated based on the intervention effect observed in the pilot study, with the LATCH Breastfeeding Assessment Tool (LATCH) used as the outcome measure. LATCH consists of five components: L (Latch), A (Audible swallowing), T (Type of nipple), C (Comfort), and H (Hold/positioning). The sample size was calculated using PASS 15.0 software, with a statistical power (1-β) of 80% and a two-tailed significance level (α) of 0.05. A two-sample t-test assuming unequal variances was employed for this calculation. The estimated sample size was N1 = N2 = 141. Considering a dropout rate of 30%, the final sample was estimated to be 402 participants.

Randomization and blingding

The research subjects were recruited by specialized outpatient nurses at the obstetric clinic and subsequently incorporated into the established pool of pregnant women. The participants were categorized based on the level of family support for breastfeeding, which was divided into three tiers: high, moderate, and none. The study statistician allocated the different types of participants in a 1:1 ratio, with odd numbers assigned to the intervention group and even numbers to the control group. The study participants were blinded to their group assignments, while the research team was aware of the group allocations. Participants were aware that different experimental conditions existed in the study, but were unaware of their specific group allocation and the detailed differences between intervention protocols. This procedure was adopted to effectively control for response bias and expectancy effects.

Intervention

As a classic positive psychological intervention, the "Three Good Things" exercise has been demonstrated to significantly enhance positive emotions, while being simple to implement and promote at the individual level²⁵. The "Three Good Things" exercise refers to individuals spending a few minutes documenting three positive events that occurred during the day. Our intervention group carried out the "three good things" intervention, and the control group carried out the “first three things” intervention, and the relevant data were collected in the form of questionnaires. Participants in the control group were asked to record the first three thoughts or events that came to mind during the day, regardless of whether these experiences were positive, negative, or neutral. Unlike the intervention group, which focused exclusively on positive events in the “Three Good Things” task, the control group did not selectively focus on positive experiences or engage in deliberate positive reappraisal. Instead, they simply documented the earliest three mental contents or daily events that arose naturally, without any requirement to emphasize positive emotions or interpret events in a positive manner. Intervention information was regularly delivered to study participants via WeChat by nurses and doctors who had received psychological qualification training. The intervention was administered every other day throughout each week. During the week, a weekly questionnaire was completed upon the third intervention session. This weekly cycle was repeated until the prenatal period. To improve participant compliance and engagement, weekly instructions and thematic prompts will be provided (Table 1). To ensure that participants complete the task on schedule, reminder messages will be sent. Participants who fail to participate for three consecutive times will be classified as dropouts. Monitor will verify the implementation of the intervention by medical staff on a weekly basis to confirm that the intervention is delivered in a standardized manner in accordance with the protocol and that eligibility criteria are strictly applied.

Table 1.

Instructions for the intervention and control groups.

Week	Topic	Instructions for the intervention group	Instructions for the control group
1	My baby	Please take a moment to recall three good things that happened today related to you and your baby, and write them down along with the reasons behind them. For example: (1). My baby moved today, which is a sign of healthy development (2). I saw a cute baby picture today, and I can’t wait to meet my own little one (3). I bought tiny clothes for my baby today—they’re so small and sweet	Please record the first three baby‑related events that come to mind today, along with the reasons why they occurred. For examples: (1). I suddenly realized how challenging it is to have a child (2). I attended a prenatal check‑up today (3). I bought some daily necessities for my baby
2	Nature and me	Please take a moment to think of three good things related to nature that happened to you today, and write them down together with the reasons behind them. For example: (1). I heard a little bird chirping on the road today—it felt like a warm hello (2). The small flower we’ve been nurturing is finally blooming, thanks to all the tender loving care we’ve given it (3). Today’s sky was a brilliant blue, with faint lines from planes streaking across	Please record the first three nature‑related events that come to mind today, along with the reasons why they occurred. For examples: (1). It is raining today, and the air feels damp and moist (2). I watered the flowers at home (3). I noticed the leaves on the trees were swaying gently in the wind
3	My family	Please take a moment to think of three beautiful things related to your family that happened to you today, and write them down together with the reasons behind them. For example: (1). My husband is incredibly thoughtful, and it makes me feel so warm and happy (2). My mom and mother-in-law specially cooked duck for me today—they always shower me with so much love and care (3). All the people I love are eagerly awaiting our baby with me, and my friends have told me to take it easy and not to worry too much	Please record the first three family-related events that come to mind today, along with the reasons why they occurred. For examples: (1). My husband and I had dinner with our parents today (2). A close friend of mine got married today (3). My family chatted with me about preparations for the baby
4	My future	Please take a moment to think of three beautiful things related to the future that happened to you today, and write them down together with the reasons behind them. For example: (1). Today I saw other people’s little kids, and I can’t wait to have my own baby this time next year! (2). The weather forecast predicts lovely weather tomorrow, and I’m really looking forward to the sunshine (3). My favorite TV series will be updated tomorrow, and I’m so excited to watch it	Please record the first three future‑related events that come to mind today, along with the reasons why they occurred. For examples: (1). We have a prenatal checkup scheduled for tomorrow (2). My mother-in-law will come to visit us tomorrow (3). I plan to organize the baby’s room this weekend
5	My life	Please take a moment to think of three beautiful things related to your daily life that happened to you today, and write them down together with the reasons behind them. For example: (1). Today I listened to my favorite songs, which brought me many warm inspirations and precious memories (2). I came across a gentle quote on my phone today: “I was born ordinary, yet I was born proud. Let us be happy, ordinary people.” (3). The emoticons my friend sent me today were so lovely; I found them really fun and amusing	Please record the first three life‑related events that come to mind today, along with the reasons why they occurred. For examples: (1). I watched videos on my phone for an hour today (2). I watched a TV series with my husband this evening (3). I tidied up my room and made it clean and tidy

Data collection

After being enrolled in the study, the participants were required to complete the baseline questionnaire first. The baseline data mainly cover educational background, occupation, annual disposable personal income, mode of delivery, corset size worn before pregnancy, current bra size, age, gestational week at entry, pre-pregnancy weight, pre-pregnancy height and pre-pregnancy BMI, household permanent residents count. Furthermore, after completing three interventions each week, a weekly questionnaire needs to be completed. The intervention was administered weekly until the patient was admitted to hospital for delivery. Upon admission, the subjects underwent prenatal blood sampling for the detection of PRL and IGF-1. Serum PRL was measured by chemiluminescent immunoassay using a Beckman Coulter UniCel DxI 800 analyzer with Beckman reagents. The inter-assay CV was < 8%. Serum IGF-1 was measured by magnetic particle chemiluminescence assay using a Snibe MAGLUMI X8 analyzer with Snibe reagents. The inter-assay CV was < 8%. The time of all blood collection was chosen at 7–8 o ‘clock in the morning. Previous studies showed that the levels of PRL and IGF-1 hormones were more stable in this time period^26,27. Given that the mode of delivery influences the secretion of the hormones being measured, venous blood samples for postpartum hormone testing are collected on the second morning post-delivery for natural births, and on the morning following two nights post-delivery for cesarean sections. One week after delivery, a postpartum questionnaire was sent to the participants via WeChat, which mainly covered the mode of delivery, emotional state, as well as breastfeeding behaviors and effects during postpartum hospitalization.

Study outcomes

Primary outcomes included lactation-related hormone secretion (PRL and IGF-I), exclusive breastfeeding behavior, the number of daily breastfeeding attempts, and neonatal formula consumption. Secondary outcome indicators included other breastfeeding behaviours during postpartum hospitalization (first breastfeeding; LATCH breastfeeding score), postpartum breastfeeding intention, postpartum emotion (PANAS positive emotion, PANAS negative emotion, Barbara Fredrickson positive emotion (BF positive emotion), anxiety, depression, resilience), and the quality of sleep of primipara.

Statistical methods

The main statistical methods included descriptive analysis, T test, rank sum test, Chi-square test, Fisher exact probability method, covariance analysis, generalized estimating equations (GEE), intermediate effect analysis, etc. In this study, mean (standard deviation) or median (interquartile distance), frequency and percentage were used to describe the measurement data. The two independent samples T test was used to compare the two groups. The rank sum test of two independent samples was used when normality was not satisfied. Covariance analysis was used for dependent variables of measurement, counting and rank data. Repeated measurements of data were analyzed using generalized estimation equations. The covariates of GEE are age and permanent household population. R 4.5.2 software was used to analyze the mediating effect between maternal PRL and IGF-I (mediating variable M) and maternal emotion (independent variable X) and exclusive breastfeeding behavior (dependent variable Y) during postpartum hospitalization. If Bootstrap is set to 5000 and the 95% Confidence Interval (95%CI) of the intermediary effect value Bootstrap does not contain 0, the intermediary effect is established. SPSS 25.0 software was used for statistical analysis of the data. All tests were two sided, with P < 0.05 considered as statistically significant.

Results

During the actual participant recruitment, a total of 405 participants were initially recruited in this study from January 2022 to August 2022. Although the entire process would not cause any harm to the participants, 108 were dropped off, and the loss of follow-up rate was 26.67%. Totally, 297 people completed this study. In the post-test questionnaire, 6 invalid questionnaires were excluded due to incorrect selection in the operational test items, resulting in 291 valid questionnaires being retained, yielding an effective rate of 97.98%. There were no statistically significant differences in baseline characteristics between the participants lost to follow-up and those who completed the study (see Appendix 3 for details).

To minimize the influence of the time to start postpartum breastfeeding on hormone secretion (see Appendix 4 for details), 204 subjects who had breastfed during the period from postpartum to the first blood collection by nurses were selected for final inclusion in this study. A total of 109 participants in the control group and 95 participants were included to verify the study hypothesis, as shown in Fig. 1.

Fig. 1.

Implementation of inclusion and exclusion of study subjects.

Baseline differences

There was a significant difference of household permanent residents count between the intervention group and the control group (χ2 = 11.56, P < 0.01). No statistically significances were shown on educational background (χ² = 0.24, P > 0.05), occupation (χ² = 2.87, P > 0.05), annual disposable personal income (χ² = 0.82, P > 0.05), mode of delivery (χ² = 0.21, P > 0.05), corset size worn before pregnancy (χ² = 3.73, P > 0.05), current bra size (χ² = 7.07, P > 0.05), age (F = 2.41, P > 0.05), gestational week at entry (F = 1.36, P > 0.05), pre-pregnancy weight (F = 0.05, P > 0.05), pre-pregnancy height (F = 1.79, P > 0.05) and pre-pregnancy BMI (F = 0.18, P > 0.05). See Tabel 2 for details.

Table 2.

Sociodemographic characteristics of the subjects (N = 204).

Basic situation		Intervention Group(N = 95)	Control Group(N = 109)	F/χ2	P
Age (years)		29.56 ± 2.93	30.23 ± 3.11	2.41	0.122
Education				0.24	0.887
	Specialized subject and the following	21	27
	Undergraduate	54	61
	Graduate or above	20	21
Occupation				2.87	0.825
	Civil servants/public institutions	14	17
	Company/business staff	40	48
	Workers	22	18
	Free employee /individual enterprises	10	18
	Farmer	3	3
	Students	1	1
	Unemployment	5	4
Household resident	Free employee /individual enterprises			11.56	0.009
	1	39	42
	2	33	28
	3	10	31
	Others	13	8
Personal disposable income per year				0.82	0.991
	80,000 yuan (including 80,000)	24	29
	8–100,000 yuan (contain 100,000 yuan)	20	19
	10–120,000 yuan (contain 120,000 yuan)	9	12
	120,000 to 150,000 yuan (including 150,000)	6	9
	150,000 to 200,000 yuan (including 200,000 yuan)	9	10
	200,000–250,00 0 yuan (including 250,000 yuan)	6	6
	More than 250,000 yuan	21	24
Gestational week before enrollment		35.81 ± 1.55	36.03 ± 1.14	1.36	0.245
Delivery way				0.21	0.648
	natural delivery	69	76
	cesarean delivery	26	33
Corset size worn before pregnancy				3.73	0.445
	A cup	22	36
	B cup	50	56
	C cup	18	13
	D cup	4	3
	E cup and above	1	1
Corset size currently worn				7.07	0.216
	A cup	3	7
	B cup	28	47
	C cup	42	36
	D cup	18	16
	E cup and above	4	3
Prepregnancy weight		55.07 ± 10.89	54.96 ± 9.61	0.05	0.942
Prepregnancy Height		160.66 ± 4.50	161.54 ± 4.88	1.79	0.183
Prepregnancy BMI		21.29 ± 3.76	21.07 ± 3.61	0.18	0.668

*Note: BMI = body mass index.

Effects of interventions on outcomes

Primary outcomes

According to the analysis of generalized estimation equation (see Table 2, Table 3 and Appendix 6 for details), there was no statistical difference in the main effect of PRL and IGF-1 levels between the intervention group and the control group after intervention (P > 0.05). Puerperal PRL level increased significantly (P < 0.001), while puerperal IGF-1 level decreased significantly (P < 0.001). The interaction effect was statistically different only for PRL (P = 0.010). There was a statistical difference between the two groups (P = 0.031), which further indicated that the intervention could effectively promote the secretion of PRL. Furthermore, an ANOVA analysis performed on participants in both groups who used formula showed that the neonatal formula consumption of intervention group (351.60 ± 349.65 ml) had a lower level than the control group (466.11 ± 464.06), with a marginally significant trend (P = 0.078). No statistically significant results were shown among the main outcome indicators, including exclusive breastfeeding behavior, average daily breastfeeding attempts, and neonatal formula consumption (See Appendix 7 for details).

Table 3.

Generalized estimating equation analysis of prolactin level of primiparas in the two groups.

Group	Prenatal (uIU/mL)	Postpartum (uIU/mL)	χ2	P
Intervention group	6847.14 ± 2640.39	8828.81 ± 4070.20	25.25	0.0013
Control group	7352.38 ± 3091.85	7976.49 ± 3074.03	3.7	0.0553
LS mean	−505.231	1024.002
95%CI	(−1326.08, 315.61)	(−222.85, 1927.50)
T/F value	−1.2141	4.712
P-value	0.2261	0.0312

¹Calculated based on T test calculated.

²based on the analysis of covariance, covariate for domestic resident population.

³calculated based on the generalized estimating equations,a covariate for resident population age, family.

*Note: LS mean = least squares mean; CI = confidence interval.

Secondary outcomes

Emotional differences between groups

In the trend of emotional change over time between the intervention group and the control group (see Figs. 2), only the resilience of the intervention group was consistently higher than that of the control group over time. According to the analysis of GEE (see Table 4 and Appendix 5 for details), compared with the control group, only resilience (higher in the intervention group than in the control group) showed a statistical difference (P = 0.033). In contrast, no statistically significant differences were observed between the two groups in terms of anxiety, depression, positive and negative affect scores from the Positive and Negative Affect Schedule (PANAS), or Barbara Fredrickson (BF).

Fig. 2.

Time trends of resilience, BF positive emotions, PANAS positive emotions and PANAS negative emotions of primipara in the two groups.

Table 4.

Generalized estimating equation analysis of resilience of primiparas in the two groups.

Group	Enrollment(T0)	Week 1(T1)	Week 2 (T2)	Week 3 (T3)	Week 4 (T4)	Postpartum (T5)	χ2	P
Intervention group	18.46 ± 1.95	17.77 ± 2.10	17.92 ± 1.92	17.80 ± 2.27	18.00 ± 1.83	15.55 ± 3.65	91.76	0.0012
Control group	18.27 ± 2.08	17.55 ± 1.97	17.20 ± 1.73	17.04 ± 2.01	17.50 ± 1.52	15.01 ± 3.49	66.56	0.0012
LS mean	0.201	0.102	0.502	0.812	0.222	0.412
95%CI	(0.36, 7.56)	(0.47, 0.68)	(0.22, 1.22)	(0.36, 1.98)	(1.43, 1.88)	(0.56, 1.38)
Wald Chi-square	0.701	0.122	1.842	1.862	0.072	0.702
P-value	0.4871	0.7252	0.1752	0.1732	0.7912	0.4042

¹Calculated based on the T test was calculated.

²Based on the generalized estimating equations, covariate for resident population age, family.

*Note: LS mean = least squares mean; CI = confidence interval.

Other secondary indicators

There were no significant differences for primipara during hospitalization about the first breastfeeding, the breastfeeding skills of primipara, perceptions of breastfeeding in primipara, intention to breastfeed, neonatal breast milk satiety, the sleep quality of primipara, and Daily number of bowel movements of the baby. But the daily urination times of the baby in intervention group (3.83 ± 1.93) had a lower level than the control group (4.41 ± 2.35), with a marginally significant trend (P = 0.056). See Table 5 and Appendix 7 for details.

Table 5.

Effects of interventions on maternal breastfeeding related outcomes.

Variables	Total	Intervention group	Control group	Statistic	P-value
Average daily Breastfeeding attempts(time)	7.36 ± 2.80	7.15 ± 2.36	7.54 ± 3.13	F = 6.70	0.414
Neonatal breast milk satiety (population(%))				χ2 = 1.25	0.263
Yes	32(15.7%)	12(12.6%)	20(18.3%)
No	172(84.3%)	83(87.4%)	89(81.7%)
Postpartum first bite feeding(population(%))				χ2 = 1.39	0.458
Breast milk	166(81.4%)	79(83.2%)	87(79.8%)
Milk powder	25(12.3%)	12(12.6%)	13(11.9%)
Others	13(6.4%)	4(4.2%)	9(8.3%)
Exclusive breastfeeding(population(%))				χ2 = 0.86	0.354
Yes	49(24.0%)	20(21.1%)	29(26.6%)
No	155(76.0%)	75(78.9%)	80(73.4%)
Whether to pump breast milk every day during hospitalization(population(%))				χ2 = 0.23	0.628
Yes	126(61.8%)	57(60.0%)	69(63.3%)
No	78(38.2%)	38(40.0%)	40(36.7%)
Daily urination times of the baby	4.14 ± 2.18	3.83 ± 1.93	4.41 ± 2.35	T = −1.92	0.056
Daily number of bowel movements of the baby	2.97 ± 1.99	3.04 ± 2.07	2.91 ± 1.92	T = 0.48	0.632
Experience breastfeeding more than imagined before birth (population(%))				χ2 = 2.65	0.449
More difficult	133(65.2%)	64(67.4%)	69(63.3%)
Equally difficult	40(19.6%)	16(16.8%)	24(22.0%)
Equally easy	12(5.9%)	4(4.2%)	8(7.3%)
More easy	19(9.3%)	11(11.6%)	8(7.3%)
LATCH	11.92 ± 1.67	12.03 ± 1.70	11.82 ± 1.64	F = 0.99	0.321
Whether sleep is affected (population(%))				χ2 = 1.77	0.413
Have no effect	2(1.0%)	2(1.8%)	0(0.0%)
Have a little effect	95(46.6%)	50(45.9%)	45(47.4%)
Have a full effect	107(52.5%)	57(52.2%)	50(52.6%)
The importance of exclusive breastfeeding (population(%))				χ2 = 3.05	0.549
Totally unimportant	1(0.5%)	0(0.0%)	1(0.9%)
Unimportant	4(2.0%)	2(2.1%)	2(1.8%)
Common	18(8.8%)	11(11.6%)	7(6.4%)
Important	83(40.7%)	40(42.1%)	43(39.4%)
Vitally important	98(48%)	42(42.2%)	56(51.4%)
Neonatal formula consumption(mL)	411.91 ± 416.63	351.60 ± 349.65	466.11 ± 464.06	F = 3.15	0.078

*Note: LATCH = LATCH breastfeeding assessment tool.

Mediation effect

We conducted mediation analyses to explore the role of PRL (M) in the relationship between resilience (X) and three distinct outcomes (Y): postpartum first bite feeding, exclusive breastfeeding, and average daily breastfeeding attempts. Missing data were handled using multiple imputation with five datasets generated via predictive mean matching to maximize available sample size and statistical power. No statistically significant mediating effects of PRL were observed across all three models (Appendix 8). Post-hoc power analyses further revealed statistical power of merely 5% for each mediation model, well below the conventional 80% threshold. Taken together, these extremely low power values, coupled with the undersized sample, suggest that the null mediation findings should be interpreted with caution, as they may reflect insufficient statistical precision rather than the true absence of the proposed mediating mechanism (resilience → PRL → breastfeeding-related outcomes).

Discussions

To our knowledge, the application of positive psychology to promote colostrum secretion represents an innovative approach in current research. The results indicated that the "Three Good Things" intervention can effectively enhance mothers’ resilience and increase their prolactin levels, although it may not promote exclusive breastfeeding behaviors. This is an extremely important and valuable discovery for research in this field.

"Three Good Things" intervention can regulate the secretion of lactation-related hormones in primiparous women. Furthermore, the trends of reduced neonatal formula consumption and fewer daily urination times in intervention group strongly support this point. Previous studies indicated that maternal adverse psychological mood was associated with a decline in maternal PRL level²¹. Although the proposed link between resilience, 5-HT neurotransmission, and PRL secretion remains partially speculative, indirect evidence from lactating models supports this biologically plausible pathway. Central 5-HT systems mediate suckling-induced PRL release, and 5-HT receptors critically regulate PRL secretion during lactation^28,29. Concurrently, elevated PRL in postpartum females suppresses HPA axis reactivity and enhances stress resilience, thereby linking PRL to adaptive emotional and behavioral outcomes^30,31. In addition, reduced chronic stress has been reported to enhance 5-HT neurotransmission and 5-HT1A auto-receptor sensitivity³². Certainly, this framework provides partial support for the hypothesized resilience-5‑HT‑PRL pathway during lactation. However, definitive confirmation will require further experimental evidence. There was no significant effect of the intervention on IGF-1 level. However, antenatal anxiety has been reported to decrease maternal IGF-1 levels²². Different from the previous studies, which focused on prenatal anxiety, the present study focuses on postpartum anxiety. The reason for the unsatisfactory results of this study may be that IGF-1 hormone is not so sensitive to changes in anxiety, and the change in its hormone level may require a longer response period. In addition, IGF-1 may retained in local tissue, rather than timely release into the blood.

"Three Good Things" intervention was effective in improving the resilience of primiparous women during hospitalization, but the effect on anxiety, depression, positive emotions, and negative emotions was not significant. However, the trend showed that positive emotion was higher, and negative emotion and anxiety were lower than the control group in postpartum period. This suggests that the "Three Good Things" intervention may have a positive effect on maternal psychological state. The results were consistent with previous studies, indicating that positive psychology intervention can reduce the effect of stress, anxiety³³, and significantly improve participants’ positive emotions²⁵. The non-significant results of this study may be due to the short duration of the intervention. As the group of participants are in gestational age of 34 weeks to 38 weeks in pregnancy, while pregnant women will be delivery in about 40 weeks. This can lead to the short duration of intervention. Future studies using the "Three Good Things" intervention could therefore appropriately extend the duration of the intervention to achieve the target effect.

Meanwhile, the results indicated that "Three Good Things" intervention had no significant effect on breastfeeding behaviors. Previous studies indicated that stress management consultation can reduce postpartum anxiety and distress, which can increase breastfeeding rate³⁴. The non-significant between-group difference in breastfeeding behavior may be attributed to the high baseline breastfeeding literacy in our sample. As this study was conducted in a baby-friendly hospital, all participants received standardized breastfeeding training and routine professional care. Such universal support may have minimized group differences and attenuated the intervention effect on breastfeeding behavior. In addition, confounders such as exposure to advertising and marketing of formula were not adjusted. Furthermore, the short duration of the intervention also constitutes a significant influencing factor. Future breastfeeding intervention studies need to consider the possible influence of the social and business environment on the outcomes.

There are some limitations should be considered when interpreting the results. Firstly, the changes of hormones during pregnancy could not be dynamically monitored. However, we tried to minimize this interference and chose to collect blood during the same time period on the first postpartum day. Secondly, due to technical limitations and the complexity of distinguishing endogenous from exogenous sources, we did not detect serum oxytocin, although this has been widely reported to be associated with lactation. Thirdly, one limitation of this study was that the sample size was not recalculated based on prolactin levels, as the biochemical detection systems and measurement units differed between the pilot and formal study hospitals, which restricted the statistical power for this outcome. Fourth, unmeasured factors including maternal diet and infant suckling frequency may act as potential confounding variables and were not fully adjusted for in the present study. Fifth, the intervention was implemented only from 34 to 38 weeks’ gestation, and this relatively short duration may further limit the generalizability of the observed results.

Conclusions

The study demonstrated that the "Three Good Things" psychology intervention was effective in enhancing the psychological resilience of primiparous women and promoting colostrum secretion. Although the psychological intervention implemented in the present study did not yield a statistically or clinically significant effect on exclusive breastfeeding behavior among the participants, this non-significant finding does not negate the value of exploring the role of emotional factors in breastfeeding outcomes. Future research should investigate the mechanisms by which emotional factors (e.g., maternal positive affect, resilience) shape breastfeeding practices. This will clarify pathways underlying intervention efficacy and provide an empirically grounded basis for developing more targeted, theory-driven interventions to support optimal breastfeeding outcomes. In addition, cognitive-behavioral interventions integrated with positive psychology, built on routine health education, could be explored to examine their impact and efficacy on breastfeeding practices.

Author contributions

All authors have made substantial contributions to the work, and the following statements accurately reflect their specific roles. All authors have read and approved the final manuscript, and agree to be accountable for all aspects of the work to ensure that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. [1] (e.g., Li Ding): Conception and design of the study; data curation (recruitment of primiparous women, collection of lactation-related data and emotional assessment data); writing—original draft. [2] (e.g., Mei He, Yingying Zhu, Xiaojin Zhang): Formal analysis (statistical analysis of data using Generalized Estimating Equations [GEE]); writing—review & editing (revising the manuscript for methodological rigor). [3] (e.g., Xiaojin Zhang): Data curation;Project administration;Visualization. [4] (e.g., Ying Zhao): Supervision; funding acquisition (securing grants for the study); writing—review & editing (finalizing the manuscript and ensuring academic integrity); correspondence for the manuscript.

Funding

Medical Scientific Research Foundation of Zhejiang Province, China (grant number 2021ZH054).

Data availability

All data supporting the findings of this study are available within the paper and its Supplementary Information.

Declarations

Competing interests

The authors declare no competing interests.

Ethics

The protocol was approved by the Ethics Committees of the Women’s Hospital School of Medicine at Zhejiang University (20190088) and the Wuyi First People’s Hospital (20210001) in accordance with the Declaration of Helsinki of the World Medical Association (WMA) and China’s Measures for the Ethical Review of Life Science and Medical Research Involving Humans.