This study adopted a cross-sectional design and was conducted across four hospitals in Shanghai, including two general hospitals and two specialized maternity hospitals, between February and July 2024. A convenience sampling approach was employed to recruit postpartum women who attended routine 42-day follow-up visits at these institutions. Eligibility criteria included: (1) age 18 years or older; (2) ability to communicate in Mandarin and comprehend the study instruments; (3) completion of a live birth; and (4) completion of the Chinese City Birth Trauma Scale (City BiTS) [29]. All eligible women, including those scoring zero, were invited to participate. In subsequent subgroup analyses, we also examined the prevalence and correlates of clinically significant symptom levels based on a total score ≥ 28, as proposed in a Brazilian validation study, while retaining the full sample for analyses of symptom severity [20]. Women were excluded if they had severe postpartum complications that prevented participation (e.g., postpartum hemorrhage requiring intensive care), significant psychiatric illness diagnosed prior to pregnancy, or a score of zero on the City BiTS.
Sample size estimation was guided by the principle that at least 5–10 participants are required per independent variable in regression analysis. With the Chinese Postpartum Social Support Scale containing 34 items, a minimum of 170 participants was required. Sample size estimation was further supported by an a priori power analysis. Based on previous studies examining the association between postpartum social support and trauma severity, we anticipated a small-to-moderate effect size (f2 = 0.08) in the final multivariable regression model. Using G*Power 3.1 for linear multiple regression with a fixed model and R [2] deviation from zero, and specifying α = 0.05, power (1 − β) = 0.80, and 10 predictors, the minimum required sample size was calculated to be 184. Allowing for a potential 20% non-response or invalid questionnaire rate, the adjusted target sample size was 230, which was achieved in the present study. All participants provided written informed consent before data collection, and the study protocol was reviewed and approved by the Ethics Committee of the Second Affiliated Hospital of Naval Medical University (Approval No. 2024SLYS6). The research adhered to the ethical principles of the Declaration of Helsinki (2013 revision).
Instruments
Symptoms indicative of childbirth-related psychological and physical distress (“postpartum birth trauma symptoms”) were measured using the City Birth Trauma Scale (City BiTS), originally developed by Ayers et al. [19] based on DSM-5 PTSD criteria and subsequently translated and validated for Chinese populations by Nie et al. [29]. The scale comprises 29 items across two symptom dimensions (birth-related and general), along with criteria for symptom duration, distress, and functional impairment. Responses are scored on a frequency scale from 0 (none) to 3 (≥ 5 times), with higher scores indicating more severe trauma symptoms. In the Chinese validation study, the total scale α was 0.934. In our sample, the total scale α was 0.93, with the birth-related symptom dimension α = 0.92 and the general symptom dimension α = 0.91. These values indicate excellent internal consistency for both instruments in this study.
Postpartum social support was assessed using the Postpartum Social Support Questionnaire (PSSQ) originally developed by Logsdon et al. [30] to measure perceived social support during the postpartum period. The Chinese version of the scale was translated, culturally adapted, and psychometrically validated by Lu et al. [31]. This version evaluates both the importance of and the actual support received in four domains: material, emotional, informational, and comparison support. Each item is rated on a 1–7 Likert scale, with higher scores indicating greater perceived importance or greater actual support received. For each domain, an “expectation gap” score was calculated as: expectation gap = importance score – actual support score. A positive gap value indicates that the perceived importance of the support exceeded the support actually received (i.e., unmet expectations), a value of zero indicates that the perceived importance and received support were equal, and a negative value indicates that the received support exceeded the perceived importance. This operationalization is consistent with the original Chinese validation of the PSSQ by Lu et al. [31]in which the importance score is interpreted as a proxy for the individual’s subjective expectation for that type of support, regardless of whether it is expected from partners, family members, or healthcare professionals. Therefore, larger positive gap values reflect a greater degree of perceived unmet need. In the validation study by Zhong et al. [31]Cronbach’s α values for the four subscales ranged from 0.893 to 0.944, with a total scale α of 0.94, indicating high internal consistency. In our sample, Cronbach’s α values for the PSQ subscales were 0.90 (material support), 0.91 (emotional support), 0.92 (informational support), and 0.89 (comparison support), with a total scale α of 0.94, confirming excellent reliability in the current study population.
Partner involvement was assessed via a structured self-report item asking participants to classify their partner’s involvement during pregnancy, childbirth, and the postpartum period as high (active participation in prenatal visits, presence at delivery, daily caregiving), medium (inconsistent or partial participation), or low (minimal presence or involvement). This classification method was adapted from existing perinatal partner-support frameworks, with operational definitions provided to participants to enhance reliability.
Data collection
Data collection occurred during scheduled 42-day postpartum clinic visits. Trained research nurses explained the study purpose and procedures to eligible participants in a private consultation room to ensure confidentiality. Demographic and obstetric data—including age, parity, education level, household registration, occupation, monthly household income, primary infant caregiver, and feeding method—were obtained via a structured questionnaire. Participants then completed the City BiTS and PSQ scales electronically using a secure QR-code-linked platform on a tablet provided by the study team. The survey platform incorporated logic checks, required responses for all items, and prevented submission of incomplete questionnaires. To minimize response bias, research staff were available to clarify questions but did not influence responses. For the purpose of statistical analysis, participant age was categorized as < 30 years or ≥ 30 years. This threshold was selected because, in Chinese obstetric practice, age ≥ 30 years for primiparous women is often associated with increased obstetric risk and corresponds to the conventional definition of “advanced maternal age” in many clinical guidelines [32]. Additionally, 30 years approximated the median age in our sample, allowing for balanced subgroup sizes. Monthly household income was dichotomized at 8,000 RMB, which is close to the 2023 Shanghai average per capita monthly disposable income reported by the National Bureau of Statistics, thus providing a contextually relevant socio-economic classification [33].
Statistical analysis
All data were analyzed using SPSS Statistics version 27.0 (IBM Corp., Armonk, NY, USA). Continuous variables were tested for normality using the Shapiro–Wilk test and presented as mean ± standard deviation (SD) if normally distributed; group comparisons were performed using independent-samples t-tests or one-way ANOVA as appropriate. Categorical variables were summarized as frequencies and percentages and compared using chi-square or Fisher’s exact tests. Pearson correlation coefficients were calculated to assess associations between trauma severity and (1) importance scores, (2) actual support scores, and (3) expectation gap scores in each support domain. Multiple linear regression analyses were performed in three steps: Model 1 examined unadjusted associations between each social support gap score and trauma severity; Model 2 adjusted for demographic covariates (age, education, household income, parity, occupation, household registration, and feeding method); Model 3 further adjusted for partner involvement. Additional models were constructed replacing support gap scores with the corresponding actual support scores to examine their independent effects. All regression results are presented with unstandardized coefficients (B), standardized coefficients (β), 95% confidence intervals (CI), t-values, and p-values, regardless of statistical significance. Categorical variables were dummy-coded, and multicollinearity was assessed using variance inflation factors (VIF < 5). Statistical significance was defined as a two-tailed p-value < 0.05.