Prevalence and associated factors of EPDS-defined depressive symptoms in fathers of preterm infants admitted to the NICU: a cross-sectional study

Abstract

Background

The admission of a preterm infant to the neonatal intensive care unit (NICU) represents a stressful event for parents and families, often precipitating negative emotions such as anxiety and sadness, and in some cases, depressive symptoms. The impact of preterm infants’ admission to the NICU within 2–3 days after birth on maternal EPDS-defined depressive symptoms has been extensively studied, but research on EPDS-defined depressive symptoms arising from this context remains scarce. This study aimed to investigate the prevalence and associated factors of EPDS-defined depressive symptoms among fathers of preterm infants shortly after the infants’ admission to the NICU.

Methods

A cross-sectional study was conducted between November 2024 and October 2025 using a convenience sampling method in a tertiary maternal and child health hospital’s neonatal intensive care unit (NICU) in Gansu Province, China. A total of 379 fathers of preterm infants were recruited as study participants. The survey was conducted at 2–3 days after NICU admission. Data were collected using baseline demographic questionnaires, the Edinburgh Postnatal Depression Scale (EPDS), the Eysenck Personality Questionnaire-Revised Short Form (EPQ-RSC), and the Family Assessment Device (FAD).

Results

In this study, we used the Edinburgh Postnatal Depression Scale (EPDS) with a cutoff score of ≥ 10 to define screen-positive depressive symptoms. The prevalence of screen-positive depressive symptoms among fathers of preterm infants at 2–3 days after NICU admission was 29.55%. The associated factors included monthly income 5,000–10,000 CNY (OR = 0.557, 95%CI:0.312–0.995), paternity leave availability (OR = 0.532, 95%CI:0.291–0.974), infant birth weight 1,500-2,499 g (OR = 0.327, 95%CI:0.117–0.916), Psychoticism (OR = 1.913, 95%CI:1.592–2.299), Extraversion(OR = 0.846, 95%CI:0.739–0.968), Neuroticism (OR = 1.507, 95%CI:1.331–1.706), moderate family functioning (OR = 2.634, 95%CI: 1.356–5.117), poor family functioning (OR = 8.939, 95%CI:4.351–18.365).

Conclusion

Among fathers of preterm infants shortly after the infants’ admission to the NICU, EPDS-defined depressive symptoms were relatively common and were associated with several paternal, infant, and family-related variables. These findings warrant further exploration in multicenter longitudinal studies.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-026-09392-8.

Introduction

Preterm birth has garnered increasing global attention. Approximately one in ten newborns worldwide was a preterm infant, with documented upward trends across multiple regions [7]. China reported over one million preterm births annually, with 6.4% absolute increase in singleton preterm births and the rise from 46.8% to 52.7% in multiple gestations between 2012 and 2018 [11]. As the leading cause of neonatal mortality, preterm birth accounted for 8 deaths per 100 preterm live births [10]. Moreover, due to systemic organ immaturity, preterm infants frequently developed multisystem complications requiring specialized care in neonatal intensive care unit (NICU) [15].

During preterm infants’ NICU hospitalization, parents often experienced complex psychological distress—including grief, guilt, despair, and anxiety—with some developing post-traumatic stress disorder (PTSD) symptoms [34]. Notably, the prevalence of EPDS-defined depressive symptoms in fathers was reported as 45.3% [4]. This may be related to fathers’ need to fulfill multiple roles—spousal supporter, primary caregiver, and medical decision-maker—while simultaneously navigating work-family conflicts during early paternal role adaptation. Several stressors, particularly medical costs, occupational pressures, and inadequate social support, have been suggested to contribute to these symptoms [24].

Previous research suggests that fathers of preterm infants may have a higher likelihood of EPDS-defined depressive symptoms than fathers of term infants [8]. A longitudinal study tracking 20 preterm infant parents from shortly after NICU admission to one-year post-discharge found that 65% of fathers were screened positive for depressive symptoms during hospitalization, and 60% remained symptomatic at follow-up [35]. Screen-positive depressive symptoms peaked within the first three postnatal days and showed a statistically significant reduction by week three, though they remained elevated compared to fathers of term infants [5]. These symptoms have been associated with impaired paternal self-efficacy and role adaptation, attenuated father-infant bonding, and possible hindered infant neurodevelopment [6]. Furthermore, screen-positive depressive symptoms may also be linked to a higher prevalence of screen-positive depressive symptoms in their spouses. Despite these potential consequences, research has predominantly focused on maternal EPDS-defined depressive symptoms in preterm contexts [22, 38], with few comparative studies of term and preterm parents [30]. Thus, less attention has been given to EPDS-defined depressive symptoms in fathers of preterm infants.

A systematic review and meta-analysis identified paternal employment status, perceived perinatal stress, partner’s psychiatric history, primiparity, and marital discord as independent predictors, whereas infant health status showed no significant association [39]. Conversely, another study demonstrated that critical health conditions of preterm infants and their receipt of treatment in the NICU after birth were associated with a higher prevalence of EPDS-defined depressive symptoms [5]. A qualitative study in Malawi identified three categories of influencing factors—paternal intrinsic factors, infant health severity, and social support adequacy—that collectively shaped caregiving competence and, in turn, paternal mental health [28]. A separate study exploring Pakistani fathers’ personality traits found a significant association with EPDS-defined depressive symptoms [1]. Furthermore, impaired family functioning and low marital satisfaction significantly predicted screen-positive depressive symptoms, although existing instruments have predominantly validated these associations in maternal populations [18]. Fathers have remained largely overlooked in this research area.

In summary, this study focused on fathers of preterm infants, given that preterm infant health status, paternal sociodemographic and personality characteristics, and family functioning have been suggested to be associated with paternal mental health. Accordingly, this study aimed to systematically investigate, across these four dimensions, the prevalence and associated factors of EPDS-defined depressive symptoms among fathers of preterm infants shortly after the infants’ admission to the NICU, with the ultimate goal of raising awareness of this issue.

Methods

Design

This is a cross-sectional study, it employed a convenience sampling approach and was conducted in the NICU of a maternal and child health hospital in Gansu Province, China. The study protocol adhered to the STROBE guidelines for reporting observational studies (supplementary 1).

Participants

Using convenience sampling, we recruited fathers of preterm infants admitted to the NICU of a maternal and child health hospital in Gansu Province, China, between November 2024 and October 2025. Inclusion criteria were: (1) fathers of preterm infants with gestational age < 37 weeks; (2) fathers of preterm infants who required NICU admission after birth; (3) fathers aged ≥ 18 years. Exclusion criteria included: (1) paternal history of psychiatric disorders, intellectual disability, drug or substance abuse, or current participation in other studies; (2) paternal refusal to participate. Removal criteria were: (1) those who withdrew due to preterm infant clinical deterioration, treatment interruption, or transfer; (2) those with invalid questionnaires.

Sample size calculation: Based on the formula for estimating population proportion n= [𝑈²_𝛼/2 × (1-)] /δ² with 𝛼=0.05, and using the reported prevalence of 33% for EPDS-defined depressive symptoms in NICU settings from previous literature (𝑈_𝛼/2 = 1.96, 𝜋=0.33, 𝛿=0.05), we estimated a required sample of 340 participants. Accounting for 10% potential invalid responses, the target sample size was set at a minimum of 378 participants. The final analysis included 379 participants.

Measures

Demographic questionnaire

A self-designed instrument collected: (1) paternal sociodemographic characteristics (ethnicity, age, educational attainment, occupation, monthly income, paternity leave status and so on); (2) infant clinical characteristics (sex, gestational age, birth weight, Apgar scores and so on).

Edinburgh Postnatal Depression Scale (EPDS)

The EPDS was originally developed by Cox [9] and subsequently validated in Chinese by Lee [23]. This 10-item scale assesses mood, anhedonia, guilt, anxiety, fear, insomnia, coping ability, sadness, crying, and self-harm. Items are scored on a 0–3 scale, with total scores ranging from 0 to 30 (0–9 = normal; ≥10 = EPDS-defined depressive symptoms). International studies report Cronbach’s α coefficients of 0.81 [27], while the Chinese version demonstrates Cronbach’s α coefficients of 0.83 [21]. In this study, the EPDS is a screening scale, not a diagnostic instrument. A cutoff score of ≥ 10 was used to identify fathers with probable screen‑positive depressive symptoms.

Eysenck Personality Questionnaire-Short Scale (EPQ-RSC)

The EPQ-RSC was revised from Eysenck’s original measure, comprising 48 items across four 12-item subscales [42]. It comprises Psychoticism (P), Extraversion (E), Neuroticism (N), and Lie Scale (L). Test-retest reliability coefficients range from 0.57 to 0.60 (P) and > 0.7 (E, N, L), demonstrating good stability in Chinese populations [33].

Family Assessment Device (FAD)

Developed by Epstein based on the McMaster Model, this 60-item tool assesses seven dimensions (problem-solving, communication, roles, affective responsiveness, affective involvement, behavior control, general functioning) [13]. Total scores range from 60 to 240 (60–120 = healthy; 121–180 = moderate; 181–240 = poor functioning). International studies report Cronbach’s α coefficients of 0.72–0.92 [37]; the Chinese version demonstrates Cronbach’s α coefficients of 0.91 [26].

Data collection

Participants were recruited through convenience sampling following approval from hospital administration and relevant departments. On the day of infant admission, researchers recruited fathers of preterm infants via telephone contact and face-to-face communication in the NICU of a maternal and child health hospital in Gansu Province, China. Fathers completed the survey at 2–3 days after NICU admission. Trained researchers explained the study objectives and questionnaire completion procedures, and obtained written informed consent. Paper questionnaires were administered on-site, with researchers providing immediate clarification for any items participants found unclear. Questionnaire completion required approximately 15–20 min, with immediate collection upon completion. Researchers verified the completeness and authenticity of responses through on-site review, excluding questionnaires with more than 10% missing data, inconsistent responses, or response biases (e.g., identical answers for all items). Regular audits by designated staff ensured data quality. Double-data entry with independent verification and logic checks maintained data accuracy. Infant clinical data were obtained through medical record review. Of the 394 fathers initially approached, 15 were excluded: 8 declined participation, 5 withdrew due to infant clinical deterioration, treatment discontinuation, or transfer, and 2 were excluded because of invalid questionnaires (> 10% missing data or contradictory responses). The final analysis included 379 participants (96.2% response rate). Figure 1 presents the participant flow diagram.

Fig. 1.

The flow diagram of participants’ recruitment process in this study

Data analysis

All statistical analyses were performed using SPSS 26.0 (IBM Corp., Armonk, NY, USA). Demographic data of preterm infants and their fathers were categorized based on previous survey findings, clinical classification criteria, and local income levels. EPDS-defined depressive symptoms and family functioning were also classified according to the scales’ established scoring criteria. Both were treated as categorical variables. Personality traits, with pre-defined dimensions based on the scale, were treated as continuous variables.

Normally distributed continuous variables were presented as mean ± standard deviation, while categorical variables were reported as frequencies and percentages. For categorical variables, chi-square (χ²) tests were used in univariate analyses. Pearson’s chi-square test was used when all expected frequencies were ≥ 5. For 2 × 2 tables, Yates’ continuity correction was used when expected frequencies were ≥ 1 and < 5 and the total sample size N ≥ 40. Fisher’s exact test was used when any expected frequency was < 1, or the total sample size N < 40, or for 2 × 2 tables when expected frequencies were < 5 but did not meet the Yates’ correction conditions described above. For R×C contingency tables, Pearson’s chi-square test was used if fewer than 20% of cells had expected frequencies < 5 and no cell had an expected frequency < 1; otherwise, Fisher’s exact test was used. For continuous variables, one-way ANOVA was used. To select candidate variables for multivariable analysis, we first included all variables with p < 0.05 in univariate analyses. This p-value-driven approach is commonly used in exploratory studies to reduce the number of predictors and avoid overfitting. However, we acknowledge that clinically important confounders might be excluded if they are not statistically significant in univariate tests.

Binary logistic regression with backward stepwise elimination (based on the likelihood ratio test, with entry and stay thresholds of α = 0.05) was then employed to construct the final multivariable model. The initial model contained all significant predictors from univariate analyses: paternal age, monthly income, paternity leave availability, infant gestational age, birth weight, 1-minute Apgar score, ventilator use, psychoticism, extraversion, neuroticism, and family functioning. Variables that became non-significant after adjustment were removed one by one. Results are expressed as odds ratios (OR) with 95% confidence intervals (CI). A two-tailed p-value < 0.05 was considered statistically significant for all analyses. Multicollinearity was assessed using variance inflation factors (VIFs) and tolerance; VIF > 5 was considered indicative of significant collinearity. The results are reported in the footnote of Table 6.

Table 6.

Multivariable logistic regression analysis of EPDS-defined depressive symptoms among fathers of premature infants (N = 379)

Independent variables	B	SE	Waldχ2	P-value	OR (95%CI)
Monthly income (CNY) (Ref: <5,000)
5,000–10,000	-0.586	0.296	3.911	0.048	0.557(0.312–0.995)
Paternity leave availability (Ref: No)
Yes	-0.631	0.309	4.179	0.041	0.532(0.291–0.974)
Birth weight (g) (Ref: <1,000)
1,500-2,499	-1.117	0.525	4.520	0.033	0.327(0.117–0.916)
Personality traits
Psychoticism	0.649	0.094	47.984	<0.001	1.913(1.592–2.299)
Extraversion	-0.167	0.069	5.900	0.015	0.846(0.739–0.968)
Neuroticism	0.410	0.063	41.817	<0.001	1.507(1.331–1.706)
Family function (Ref: Good 60–120 points)
Moderate (121–180 points)	0.969	0.339	8.181	0.004	2.634(1.356–5.117)
Poor (181–240 points)	2.190	0.367	35.557	<0.001	8.939(4.351–18.365)

Exchange rate on December 2, 2025, 1 CNY ≈ 0.1413 USD. Monthly income refers to the father’s individual monthly income(in CNY).CI: confidence interval; bold: Statistically significant differences. VIF values for the predictors in the final model ranged from 1.058 to 1.935, indicating no substantial multicollinearity

Results

Demographic characteristics

The study comprised 379 fathers of preterm infants. Demographic characteristics revealed that the majority were of Han ethnicity (n = 313, 82.59%), with most participants aged 30–40 years (n = 252, 66.49%). Educational attainment data showed that over three-quarters (77.84%) had completed high school or higher education. Other occupations (manual/agricultural work) were reported by 39.05% of fathers, while more than half (53.83%) reported monthly incomes exceeding 5000 CNY. Among the corresponding 379 preterm infants, males predominated (n = 214, 56.46%) (Tables 1 and 2).

Table 1.

summarizes the general demographic and clinical characteristics of the fathers of preterm infants (N = 379)

Variable	Category	Frequency (n)	Proportion (%)
Ethnicity	Han	313	82.59%
	Ethnic minorities	66	17.41%
Age category	< 30 Years of age	103	27.18%
	30–40 Years of age	252	66.49%
	40 + Years of age	24	6.33%
Education Level	Junior high school or below	84	22.16%
	Senior high/vocational school	114	30.10%
	Bachelor’s degree	171	45.12%
	Master’s degree or higher	10	2.64%
Occupation	Government/institutional employees	130	34.30%
	Commercial/service sector staff	36	9.50%
	Healthcare professionals	6	1.58%
	Self-employed individuals	59	15.57%
	Other (manual/agricultural work)	148	39.05%
Monthly income (CNY)	< 5,000	175	46.17%
	5,000–10,000	167	44.06%
	> 10,000	37	9.76%
Planned pregnancy	Yes	297	78.36%
	No	82	21.64%
Prenatal companionship duration	< 3 months	38	10.03%
	3–6 months	77	20.32%
	Full prenatal companionship	264	69.65%
Paternity leave availability	Yes	176	46.44%
	No	203	53.56%
Experience in preterm infant	Yes	17	4.49%
care	No	362	95.51%

Exchange rate on December 2, 2025, 1 CNY ≈ 0.1413 USD. Monthly income refers to the father’s individual monthly income(in CNY)

Table 2.

summarizes the general demographic and clinical characteristics of the preterm infants (N = 379)

Variable	Category	Frequency (n)	Proportion (%)
Sex	Male	214	56.46%
	Female	165	43.54%
Gestational age (weeks)	< 28 + 0	24	6.33%
	28 + 0–33 + 6	215	56.73%
	34 + 0–36 + 6	140	36.94%
Birth weight (g)	< 1,000	36	9.50%
	1,000–1,499	102	26.91%
	1,500-2,499	203	53.56%
	2,500-4,000	38	10.03%
Apgar score (1 min)	0–3	8	2.11%
	4–7	100	26.39%
	8–10	271	71.50%
Apgar score (5 min)	0–3	1	0.26%
	4–7	28	7.39%
	8–10	350	92.35%
Apgar score (10 min)	0–3	0	0.00%
	4–7	12	3.17%
	8–10	367	96.83%
Ventilator support	None	72	19.00%
	Non-invasive ventilator	211	55.67%
	Sequential invasive-noninvasive ventilation	96	25.33%
Congenital malformation	Yes	4	1.06%
	No	375	98.94%

Univariate analysis of general demographic data of preterm infants and their fathers

At 2–3 days after NICU admission, 112 fathers of preterm infants had EPDS scores of ≥ 10, corresponding to a prevalence of screen-positive depressive symptoms of 29.55%. Chi-square (χ²) tests (including Pearson’s χ² test and Yates’ continuity correction) and Fisher’s exact tests revealed significant associations between several demographic/clinical factors and EPDS-defined depressive symptoms. Specifically, paternal age, monthly income, paternity leave availability, along with infant characteristics including gestational age, birth weight, 1-minute Apgar score, and ventilator use showed significant associations with screen‑positive depressive symptoms (p < 0.05) (Table 3).

Table 3.

Univariate analysis of EPDS-defined depressive symptoms among fathers of preterm infants (N = 379)

Variable	Fathers with EPDS-defined depressive symptoms (N = 112)	Fathers without EPDS-defined depressive symptoms (N = 267)	Statistics	P -value
Preterm infants’ fathers
Ethnicity
Han	93(83.04%)	220(82.40%)	0.022a	0.881
Ethnic minorities	19(16.96%)	47(17.60%)
Age category
<30 Years of age	43(38.39%)	60(22.47%)
30–40 Years of age	62(55.36%)	190(71.16%)	10.325a	0.006
40 + Years of age	7(6.25%)	17(6.37%)
Education Level
Junior high school or below	21(18.75%)	63(23.60%)
Senior high/vocational school	37(33.04%)	77(28.84%)	1.305a	0.728
Bachelor’s degree	51(45.54%)	120(44.94%)
Master’s degree or higher	3(2.68%)	7(2.62%)
Occupation
Government/institutional employees	40(35.71%)	90(33.71%)
Commercial/service sector staff	11(9.82%)	25(9.36%)
Healthcare professionals	2(1.79%)	4(1.50%)	2.902a	0.574
Self-employed individuals	12(10.71%)	47(17.60%)
Other (manual/agricultural work)	47(41.96%)	101(37.83%)
Monthly income (CNY)
< 5,000	66(58.93%)	109(40.82%)
5,000-10,000	37（33.04%）	130（48.69%）	10.474a	0.005
>10,000	9（8.04%）	28（10.49%）
Planned pregnancy
Yes	87（77.68%）	210（78.65%）	0.044a	0.834
No	25（22.32%）	57（21.35%）
Prenatal companionship duration
<3 months	16（14.29%）	22（8.24%）
3-6 months	23（20.54%）	54（20.22%）	3.338a	0.188
Full prenatal companionship	73（65.18%）	191（71.54%）
Paternity leave availability
Yes	42（37.50%）	134（50.19%）	5.106a	0.024
No	70（62.50%）	133（49.81%）
Experience in preterm infant care
Yes	2（1.79%）	15（5.62%）	2.705a	0.100
No	110（98.21%）	252（94.38%）
Preterm infants
Sex
Male	63（56.25%）	151（56.55%）	0.003a	0.957
Female	49（43.75%）	116（43.45%）
Gestational age (weeks)
<28+0	13(11.61%)	11(4.12%)
28+0-33+6	75(66.96%)	140(52.43%)	20.276a	＜ 0.001
34+0-36+6	24(21.43%)	116(43.45%)
Birth weight (g)
<1,000	21(18.75%)	15(5.62%)
1,000-1,499	38(33.93%)	64(23.97%)	23.867a	＜ 0.001
1,500-2,499	45(40.18%)	158(59.18%)
2,500-4,000	8(7.14%)	30(11.24%)
Apgar score (1 min)
0-3	5(4.46%)	3(1.12%)
4-7	38(33.93%)	62(23.22%)	9.778a	0.008
8-10	69(61.61%)	202(75.66%)
Apgar score (5 min)
0-3	0(0.00%)	1(0.37%)
4-7	11(9.82%)	17(6.37%)	1.863b	0.495
8-10	101(90.18%)	249(93.26%)
Apgar score (10 min)
0-3	0(0.00%)	0(0.00%)
4-7	6(5.36%)	6(2.25%)	1.578a	0.209
8-10	106(94.64%)	261(97.75%)
Ventilator support
None	10(8.93%)	62(23.22%)
Non-invasive ventilator	59(52.68%)	152(56.93%)	19.450a	＜ 0.001
Sequential invasive-noninvasive ventilation	43(38.39%)	53(19.85%)
Congenital malformation
Yes	2(1.79%)	2(0.75%)	0.123a	0.726
No	110(98.21%)	265(99.25%)

a = χ² test (Pearson's χ² test for all expected frequencies ≥5; Yates' continuity correction for 2×2 tables with 1≤expected frequencies <5 and N ≥40); b = Fisher's exact test (for expected frequency <1 or N <40, or any 2×2 table with expected frequency <5 that does not meet the Yates' correction conditions). For R×C tables: χ² test is used if <20% of cells have expected frequencies <5 and none have expected frequency <1; otherwise Fisher's exact test. bold: Statistically significant differences

Univariate analysis of personality dimensions and family functioning

One-way ANOVA demonstrated significant between-group differences in personality dimensions between fathers with and without screen-positive depressive symptoms: Psychoticism (F = 146.679, p < 0.001), Extraversion (F = 99.349, p < 0.001), and Neuroticism (F = 286.987, p < 0.001). Fathers with screen-positive symptoms showed significantly higher Psychoticism (5.56 ± 2.07 vs. 2.94 ± 1.49) and Neuroticism (7.25 ± 1.65 vs. 3.14 ± 3.03) scores, but lower Extraversion scores (6.48 ± 1.88 vs. 8.88 ± 2.64) (p < 0.001). No significant difference was observed in Lie scale (L) scores (p = 0.135>0.05) (Table 4).

Table 4.

presents the scores of personality traits among fathers of premature infants in the NICU (N = 379)

Category	Scores of fathers with EPDS-defined depressive symptoms (mean ± SD)	Scores of fathers without EPDS-defined depressive symptoms (mean ± SD)	F	P-value
Psychoticism	5.56 ± 2.07	2.94 ± 1.49	146.679	<0.001
Extraversion	6.48 ± 1.88	8.88 ± 2.64	99.349	<0.001
Neuroticism	7.25 ± 1.65	3.14 ± 3.03	286.987	<0.001
Lie	6.24 ± 2.36	6.65 ± 2.72	2.252	0.135

Regarding family functioning, participants were categorized into good (60–120 points; n = 105, 27.70%), moderate (121–180 points; n = 188, 49.60%), and poor (181–240 points; n = 86, 22.69%) functioning groups. A significant association emerged between family functioning levels and the prevalence of screen-positive depressive symptoms (χ² = 43.892, p < 0.001), with rates increasing across worsening function categories (Table 5).

Table 5.

Comparison of the prevalence of EPDS-defined depressive symptoms among fathers of premature infants with different levels of family function (N = 379)

Category	Fathers with EPDS-defined depressive symptoms (N = 112)		Fathers without EPDS-defined depressive symptoms (N = 267)
Good family function (60–120 points)	13(11.61%)	92(34.46%)
Moderate family function (121–180 points)	51(45.54%)	137(51.31%)
Poor family function (181–240 points)	48(42.86%)	38(14.23%)

Overall comparison across three groups: χ² = 43.892, p < 0.001

Multivariate logistic regression analysis

Using an EPDS cutoff score of ≥ 10 to define screen-positive depressive symptoms as the dependent variable, we constructed a multivariable logistic regression model. Given our sample size and the number of candidate variables, stepwise elimination was used to avoid model overfitting. Specifically, the initial model included all variables that were statistically significant in univariate analyses (p < 0.05): paternal age, monthly income, paternity leave availability, infant gestational age, birth weight, 1-minute Apgar score, ventilator use, psychoticism, extraversion, neuroticism, and family functioning.

Through backward stepwise elimination (likelihood ratio test, removal threshold p > 0.05), four variables were excluded from the final model because their effect estimates became non-significant after adjusting for other covariates: paternal age, infant gestational age, 1-minute Apgar score, and ventilator use. The final multivariable model included the following variables (P < 0.05): monthly income, paternity leave availability, infant birth weight, psychoticism, extraversion, neuroticism, and family functioning.

The final model revealed the following associated factors: monthly income 5,000–10,000 CNY(OR = 0.557, 95%CI:0.312–0.995), paternity leave availability (OR = 0.532, 95%CI:0.291–0.974), infant birth weight 1,500-2,499 g (OR = 0.327, 95%CI:0.117–0.916), psychoticism (OR = 1.913, 95%CI:1.592–2.299), extraversion (OR = 0.846, 95%CI:0.739–0.968), neuroticism (OR = 1.507, 95%CI:1.331–1.706), moderate family functioning (OR = 2.634, 95%CI: 1.356–5.117), poor family functioning (OR = 8.939, 95%CI:4.351–18.365) (Table 6). VIF values for all predictors ranged from 1.058 to 1.935, well below the conservative threshold of 5, indicating no substantial multicollinearity.

Discussion

The study showed that among 379 fathers of preterm infants, within 2–3 days after their preterm infants’ admission to the NICU, 112 had EPDS scores of ≥ 10, corresponding to a prevalence of screen-positive depressive symptoms of 29.55%. These screen-positive depressive symptoms underscored the critical need for healthcare providers to address paternal mental health in NICU settings [2]. Comparative analysis showed our outcome exceeds reports from the study [14]. This difference may be partly attributable to a stressful situation arising from the combination of high medical costs associated with preterm infants’ young gestational age and low birth weight, and the difficulty that fathers with low income face in balancing these high medical expenses against their limited financial resources. Conversely, our findings were lower than those reported [35], possibly reflecting methodological variations in sampling or assessment timing. It should also be noted that fathers were excluded because their infants experienced clinical deterioration or transfer; these fathers may have experienced higher levels of distress, and their exclusion could have led to an underestimation of the true prevalence of EPDS-defined depressive symptoms. A recent multinational meta-analysis documented a temporal decline in the prevalence of EPDS-defined depressive symptoms from 13% in year 1 to 11% in years 2–5 post-NICU, highlighting the importance of monitoring duration [29]. Given the predominance of cross-sectional studies in China, we recommend conducting multicenter, large-sample longitudinal studies to investigate the dynamic changes in the prevalence of these symptoms and their associated factors.

Our study found that a birth weight of 1,500–2,499 g in preterm infants was associated with a lower prevalence of EPDS-defined depressive symptoms. Data analysis showed that, compared with fathers of infants with very low birth weight (< 1,500 g), those in this birth weight group had a significantly lower prevalence of EPDS-defined depressive symptoms. This aligns with the findings [17] which demonstrate an inverse relationship between birth weight and paternal psychological distress. This finding suggested that clinicians may consider paying particular attention to the mental health status of fathers of preterm infants with birth weight below 1,500 g, while recognizing that causal inferences cannot be drawn from this cross-sectional study. Notably, clinical indicators such as gestational age, Apgar scores, and ventilator use were associated with screen-positive depressive symptoms in univariate analysis but not after multivariable adjustment. This discrepancy with those reported [20] but consistency with the findings [39] in China. These findings suggest that cultural differences may influence the psychological responses of fathers, a possibility that warrants further examination.

Although the study has identified associations between screen-positive depressive symptoms and factors such as age, education level, occupation and time spent at home [1], our study did not find these factors (age, education level, occupation type, or time spent accompanying the mother during pregnancy) to be associated with screen-positive depressive symptoms. This discrepancy may be attributable to differences in personality traits. In the present study, fathers with a monthly income of 5,000–10,000 CNY had lower prevalence of screen-positive depressive symptoms than those with a monthly income of < 5,000 CNY. This finding is consistent with studies [3, 31, 40]. However, the mechanisms underlying this association remain unclear, and the relationship between income levels and paternal distress warrants further investigation.

The present study found that unplanned pregnancy and lack of experience in preterm infant care were not associated with screen-positive depressive symptoms. However, access to paternity leave was an associated factor. It was consistent with the results of previous studies [19, 32]. Under China’s current parental leave policies, fathers either receive no paternity leave or only brief periods (typically 7–15 days in most regions). Whether this would weaken fathers’ ability to participate in preterm infant care and consequently affect their mental health warrants further examination.

Consistent with Eysenck’s Personality Theory [36], our study demonstrated significant associations between paternal personality traits and screen-positive depressive symptoms. Higher psychoticism(P) and neuroticism(N) scores were significantly associated with higher EPDS-defined depressive symptoms, while higher extraversion(E) scores were associated with lower EPDS-defined depressive symptoms. No significant association was observed for lie scale (L) scores. It has been suggested that individuals with high psychoticism may experience heightened loneliness and paranoid ideation, while those with elevated neuroticism exhibit increased sensitivity to negative stimuli; conversely, extraverted individuals typically possess more adaptive emotion-regulation strategies. These findings align with multiple studies [1, 16]. However, because personality traits and EPDS-defined depressive symptoms were assessed at the same time point, and EPDS-defined depressive symptoms may also influence how fathers report their personality traits, the direction of these associations warrants further investigation.

Regarding family functioning, our study found that moderate and poorer family functioning were associated with a higher prevalence of EPDS-defined depressive symptoms. Because family functioning and EPDS-defined depressive symptoms were measured concurrently, it is also possible that fathers with more EPDS-defined depressive symptoms perceive their family functioning more negatively. Longitudinal studies are needed to clarify the direction of this association. These findings are consistent with previous studies [12, 41], suggesting that optimal family functioning and partner support may serve as important psychosocial resources. Although existing research has predominantly examined the association between family functioning and EPDS-defined depressive symptoms in mothers [18, 25], future studies should pay more attention to fathers and also investigate the bidirectional relationship between EPDS-defined depressive symptoms and family functioning.

Throughout the Discussion, it is important to note that several potentially important factors (e.g., maternal mental health, marital relationship, social support, perceived stress, and detailed neonatal severity indicators) were not measured in this study. Therefore, the observed associations may be subject to residual confounding, and the interpretations offered are tentative pending replication in studies with more comprehensive variable coverage. Furthermore, it should be acknowledged that this study only screened for EPDS-defined depressive symptoms among fathers of preterm infants at 2–3 days after NICU admission, and the EPDS cannot be used for diagnosis. Therefore, the investigators only provided reassurance to fathers with high total EPDS scores or a positive response on item 10, without establishing a clear referral protocol. EPDS-defined depressive symptoms in fathers require subsequent long-term follow-up, with formal psychological support and a clear referral protocol to be provided when necessary.

Limitation

This study has several limitations. First, the cross-sectional design precludes determination of directionality; reverse causality cannot be ruled out. Second, the EPDS was used as a screening rather than a diagnostic tool. Third, excluding fathers who refused to participate or whose infants experienced clinical deterioration or transfer may have introduced selection bias. Fourth, several potentially important factors (e.g., maternal mental health, marital relationship, social support, perceived stress, and detailed neonatal severity indicators) were not measured in this study. Fifth, despite minimal missing data (< 4%) and adherence to established classification criteria, sensitivity analyses could not be performed because the core variable (EPDS-defined depressive symptoms) was not collected in these cases. Future multicenter longitudinal studies with larger sample sizes and sensitivity analyses are needed to confirm these associations. Sixth, although the EPDS cutoff of ≥ 10 was chosen based on its good reliability and validity, the robustness of this cutoff and the p-value-based model selection strategy are limited, and future validation is required. Seventh, the generalizability is limited due to the single-center design. Finally, for fathers with high total EPDS scores or a positive response on Item 10, the researchers only provided verbal reassurance, such as listening, explanation, and encouragement. No ad hoc safety confirmation or consultation with physicians, psychologists, or clinical staff was conducted, and no formal referral protocol or clinical safety net was in place. This should be acknowledged as an ethical and procedural limitation. Future studies should establish a clear safety assessment, referral, and follow-up protocol for fathers with possible self-harm risk.

Conclusion

Among fathers of preterm infants shortly after the infants’ admission to the NICU, EPDS-defined depressive symptoms were relatively common and were associated with several paternal, infant, and family-related variables. These findings warrant further exploration in multicenter longitudinal studies.

Abbreviations

NICU

Neonatal Intensive Care Units

EPDS

Edinburgh Postnatal Depression Scale

EPQ-RSC

Eysenck Personality Questionnaire-Short Scale

P

Psychoticism

E

Extraversion

N

Neuroticism

L

Lie

FAD

Family Assessment Device

Authors’ contributions

Shiyu Li : Research design, data collection, data analysis, drafting the initial manuscript, and manuscript revision.Yunyun Ma : Data collection, data curation, and supervision.Taomin Wu: Data collection and data curation.Jiaqi Chen: Data collection and data curation.Zhen’e Zhu: Data analysis and supervision.Jie Huang: Data analysis.Xiaohua Ding: Research design, revising the initial draft, supervision, and project administration.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Data availability

All the data, transcripts and supporting documents used for this current study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This research was conducted in compliance with the Helsinki Declaration. All participants provided written informed consent after receiving comprehensive information regarding the study protocol, potential risks, and rights protection. The study complied with fundamental ethical principles including voluntariness, confidentiality, justice, beneficence/non-maleficence, and respect for persons throughout all research procedures. This study strictly adhered to medical research ethics guidelines. It received approval from the Institutional Review Board (IRB) of Gansu Provincial Maternity and Child-care Hospital at a convened meeting on October 15, 2024 (the ethical approval document was issued on April 17, 2025, with number (2025) GSFY-IRB-57). The study period was from November 2024 to October 2025, and all participant recruitment and data collection began only after the ethical approval (October 15, 2024).

Consent for publication

It is not applicable to this study. All authors read and approved the final manuscript.

Competing interests

The authors declare that they have no conflicts of interest regarding competing financial interests or personal relationships that could have influenced the work reported in this paper.