From April to September 2022, a convenient sampling method was adopted to select women who gave birth in a third-class A hospital in Guangzhou, Guangdong Province, China, and met the inclusion and exclusion criteria. The inclusion criteria were 28 to 42 weeks of gestation; voluntary intraspinal anesthesia delivery after physician evaluation and communication with the patient; and voluntary participation in the study. The exclusion criteria were prenatal patients with urinary disease and patients with mental disorders or who were unable to communicate. The elimination criteria were conversion to cesarean section during delivery; and wanting to withdraw from the investigator for any reason.

Path analysis requires calculation based on the sample size of no less than 150 cases and 10 to 20 occurrences of the observed variables [6]. Considering a 15% shedding rate, the sample size is about 357. In this study, 372 questionnaires were distributed, and 333 valid questionnaires were collected, with an effective recovery rate of 89.52%.

The general data questionnaire was designed to include nationality, educational background, marital status, residence, occupation and medical payment method. The questionnaire on the influencing factors of PUR after intraspinal anesthesia was obtained from the previous research results of the research group. It was designed based on the Delphi method and the theory of displeasure symptoms, the influencing factors of PUR included physiological, psychological, and environmental factors [7]. After two rounds of expert correspondence, the first level item was added: medical intervention factor; as this study was a single-center study, all women faced the same medical environment after delivery and ignored the first-level item environmental factor. The modified influencing factors questionnaire included a total of three first-level items and 31 second-level entries. First-level item physiological factors were Body Mass Index (BMI) before pregnancy, weight gain during pregnancy, age, number of births, prenatal activity, pregnancy complications, the second stage of labor time, total labor time, degrees of perineal laceration, neonatal birth weight, postpartum pain level, history of urinary retention, degrees of perineal edema, and degrees of hemorrhoid edema. First-level item medical intervention factors were prenatal bed urination training, prenatal pelvic floor muscle exercise, number of catheterizations during labor, free position delivery application, oxytocin, water sac labor induction, manual rotation of the fetal head, episiotomy, forceps delivery, vacuum extractor of the fetal head, analgesic duration, and use of antispasmodic drugs during childbirth. First-level item psychological factors were depression, anxiety, psychological support, the change of postnatal urination posture, and the mastery of PUR-related knowledge.

The relevant indicators are defined as follows:

Body Mass Index before pregnancy classification was: low weight, BMI <18.5; normal weight, 18.5 to 24; overweight, 24–28; and obesity, BMI ≥28 [8]. Weight gain during pregnancy was divided according to maternal BMI, 11.0 to 16.0 kg for low weight; 8.0 to 14.0 kg for normal weight; 7.0 to 11.0 kg for overweight; and 5.0 to 9.0 kg for obesity [8]. Maternal body weight gain in the range is recorded as normal body weight gain during pregnancy, less than the range of increase range as insufficient body weight gain during pregnancy, and more than the range of increase range as excessive body weight gain during pregnancy.

Degrees of perineal laceration [9] were degree I, perineal skin and (or) vaginal mucosa damage; degree II, with perineal muscle damage; degree III, injury involving the anal sphincter; degree IV, damage to the internal and external anal sphincter and involvement of the rectal mucosa.

Degrees of perineal edema [10] were no edema; degree I, mild perineal swelling, skin pattern; degree II, perineal swelling to shiny skin, skin pattern disappeared; and degree III, perineal swelling to clear skin, and swelling of the peripheral labia.

Degrees of hemorrhoid edema [11]: Degree I, no edema; degree II, mild, mild edema, can be incorporated into the anus; degree III, moderate, obvious edema, pain; degree IV, severe, obvious edema, pain, with erosion and necrosis of internal hemorrhoids.

The postpartum pain level was assessed by the simple pain numerical rating scale (NRS) with 0–10 points. Patients were required to choose the number best representing their pain intensity from 0 to 10 points, with 0 painless pain, 1–3 indicating mild pain, 4–6 moderate pain, 7–9 severe pain, and 10 excruciating pain.

Maternal depression and anxiety were assessed using the Hospital Anxiety Depression Scale (HADS), the scale established in [12] by Zigmond et al. Cronbach's α coefficients of the Chinese subscale for anxiety and depression were 0.762 and 0.787 respectively, showing good reliability. Between the two subscales, 0–7 was classified as no depression or anxiety, 8–10 was classified as suspected depression or anxiety symptoms, and 11–21 was classified as certain depression or anxiety symptoms [13].

The prenatal activity of women used the physical activity questionnaire during pregnancy (PPAQ) proposed by Chasan-Taber et al. in [14]. The Chinese version of the questionnaire included 31 items, the content validity was 0.940, and the retest reliability was 0.944 [15]. According to the Canadian guidelines for physical activity during pregnancy, moderate physical activity for at least 150 min per week was recommended as the standard of physical activity during pregnancy [16].

The mastery of PUR-related knowledge was evaluated by self-designed questions. When puerperae thought that they had PUR knowledge and knew the importance of early postpartum urination, they needed to answer three multiple choice questions, which covered the definition, influencing factors, and preventive measures of PUR. Puerperae who had PUR-related knowledge were recorded as having 0 points, those who did have PUR-related knowledge, but got 0 correct, were recorded as having 1 point, if they got 1 correct this was recorded as 2 points, and so on; when 3 questions were all correct, this was recorded as 4 points.

In this study, a portable, non-invasive transabdominal bladder ultrasound device (Bladderscan® PBSV4.1; Mianyang Meike, China) was used to scan the bladder of a woman immediately after her first postpartum urination. The residual urinary volume after the first spontaneous urination was ≥150 ml and was recorded as PUR.

This study was approved by the ethical review body of the Third Affiliated Hospital of Guangzhou Medical University. Before data collection, consent was obtained from the relevant departments, and all pregnant women participating in the survey signed written informed consent.

Considering the energy and cooperation of the puerperae, the questionnaire was divided into two parts. One part was filled in by the puerperae, and the data included were those not available from the medical data, such as postpartum pain level, psychological support, prenatal activity, etc. The other part was filled in by the investigator, and included data that can be obtained directly from the maternal medical data, such as whether there is an episiotomy, the duration of the second stage of labor, etc. Within 6 h of delivery, puerperae who voluntarily participated in were surveyed; the purpose, significance, and questionnaire filling method were explained; and the questionnaires were filled in independently. If needed the puerpera was assisted by the investigator and the questionnaires were recalled on the spot. If there were any omissions, they were completed immediately.

Data processing was performed using SPSS25.0, and all continuous variables were converted into categorical variables before analysis. Using the frequency and composition ratio description, categorical data were compared using the Chi-squared test and nonparametric test, and p<0.05 was considered a statistically significant difference.The path analysis model was established and modified using AMOS24.0, and estimated using the maximum likelihood method, when the Chi-squared degree of freedom ratio (χ2/df) <2, comparing the fitting index (CFI) and the Tucker–Lewis index (TLI) >0.90, the root mean square error of approximation (RMSEA) <0.05, and was considered a good model fit [17].