Background: Postpartum depression (PPD) is a serious public health problem. Building a predictive model for PPD using data during pregnancy can facilitate earlier identification and intervention. Objective: The aims of this study are to compare the effects of four different machine learning models using data during pregnancy to predict PPD and explore which factors in the model are the most important for PPD prediction. Methods: Information on the pregnancy period from a cohort of 508 women, including demographics, social environmental factors, and mental health, was used as predictors in the models. The Edinburgh Postnatal Depression Scale score within 42 days after delivery was used as the outcome indicator. Using two feature selection methods (expert consultation and random forest-based filter feature selection [FFS-RF]) and two algorithms (support vector machine [SVM] and random forest [RF]), we developed four different machine learning PPD prediction models and compared their prediction effects. Results: There was no significant difference in the effectiveness of the two feature selection methods in terms of model prediction performance, but 10 fewer factors were selected with the FFS-RF than with the expert consultation method. The model based on SVM and FFS-RF had the best prediction effects (sensitivity=0.69, area under the curve=0.78). In the feature importance ranking output by the RF algorithm, psychological elasticity, depression during the third trimester, and income level were the most important predictors. Conclusions: In contrast to the expert consultation method, FFS-RF was important in dimension reduction. When the sample size is small, the SVM algorithm is suitable for predicting PPD. In the prevention of PPD, more attention should be paid to the psychological resilience of mothers.

Schizophrenia is a leading cause of disability, and a shift from facility- to community-based care has been proposed to meet the resource challenges of mental healthcare in low- and middle-income countries. We hypothesized that the addition of mobile texting would improve schizophrenia care in a resource-poor community setting compared with a community-based free-medicine program alone. In this 2-arm randomized controlled trial, 278 community-dwelling villagers (patient participants) were randomly selected from people with schizophrenia from 9 townships of Hunan, China, and were randomized 1:1 into 2 groups. The program participants were recruited between May 1, 2015, and August 31, 2015, and the intervention and follow-up took place between December 15, 2015, and July 1, 2016. Baseline characteristics of the 2 groups were similar. The patients were on average 46 years of age, had 7 years of education, had a duration of schizophrenia of 18 years with minimal to mild symptoms and nearly one-fifth loss of functioning, and were mostly living with family (95%) and had low incomes. Both the intervention and the control groups received a nationwide community-based mental health program that provided free antipsychotic medications. The patient participants in the intervention group also received LEAN (Lay health supporters, E-platform, Award, and iNtegration), a program that featured recruitment of a lay health supporter and text messages for medication reminders, health education, monitoring of early signs of relapses, and facilitated linkage to primary healthcare. The primary outcome was medication adherence (proportion of dosages taken) assessed by 2 unannounced home-based pill counts 30 days apart at the 6-month endpoint. The secondary and other outcomes included patient symptoms, functioning, relapses, re-hospitalizations, death for any reason, wandering away without notifying anyone, violence against others, damaging goods, and suicide. Intent-to-treat analysis was used. Missing data were handled with multiple imputations. In total, 271 out of 278 patient participants were successfully followed up for outcome assessment. Medication adherence was 0.48 in the control group and 0.61 in the intervention group (adjusted mean difference [AMD] 0.12 [95% CI 0.03 to 0.22]; p = 0.013; effect size 0.38). Among secondary and other outcomes we noted substantial reduction in the risk of relapse (26 [21.7%] of 120 interventional participants versus 40 [34.2%] of 117 controls; relative risk 0.63 [95% CI 0.42 to 0.97]; number needed to treat [NNT] 8.0) and re-hospitalization (9 [7.3%] of 123 interventional participants versus 25 [20.5%] of 122 controls; relative risk 0.36 [95% CI 0.17 to 0.73]; NNT 7.6). The program showed no statistical difference in all other outcomes. During the course of the program, 2 participants in the intervention group and 1 in the control group died. The limitations of the study include its lack of a full economic analysis, lack of individual tailoring of the text messages, the relatively short 6-month follow-up, and the generalizability constraint of the Chinese context. The addition of texting to patients and their lay health supporters in a resource-poor community setting was more effective than a free-medicine program alone in improving medication adherence and reducing relapses and re-hospitalizations. Future studies may test the effectiveness of customization of the texting to individual patients. Chinese Clinical Trial Registry ChiCTR-ICR-15006053.

We aimed to investigate the coping styles of Chinese pregnant women, identify factors associated with coping and further explore the effect of coping during pregnancy on postpartum depression. A longitudinal study was performed from early pregnancy to six-week postpartum. A total of 1126 women were recruited by convenience sampling and participants who completed eight questionnaires at four time points were included (three self-developed questionnaires, Coping Style Questionnaire, Generalized Anxiety Disorder-7, Brief Resilience Scale, Rosenberg Self-esteem Scale, Edinburgh Postnatal Depression Scale) (n = 615). Linear regression analyses were used to identify the possible factors for coping and their association with postpartum depression. The mean scores of positive coping and negative coping were 2.03 and 1.21, respectively. Women with a higher educational level scored higher on both positive and negative coping in pregnancy. Resilience was associated with both positive and negative coping, while self-esteem only related to positive coping (p < 0.05). Postpartum depression was associated with both positive and negative coping (p < 0.05). The women in our study reported using positive coping styles more than negative coping antenatally. Positive and negative coping behaviors could be used simultaneously. Increasing self-esteem and resilience antenatally might promote more positive coping and further reduce the occurrence of postpartum depression.

Background. Postpartum depression (PPD) is common after childbirth. Previous reviews on the prevalence of PPD have mainly included results that relied on screening instruments or a mixture of such instruments and diagnostic interviews. In this study, we aimed to assess the prevalence of PPD based exclusively on studies using diagnostic interviews, as they provide the most reliable and valid approach for defining “caseness.” Methods. Using PubMed, Web of Science, Cochrane Library, Embase, CNKI, WANFANG DATA, and CBM up to September 18, 2022, we searched for original articles reporting data that could be used to calculate the prevalence of PPD based on diagnostic interviews. A random-effect meta-analysis model was then used to estimate the pooled prevalence. In addition, we assessed quality, heterogeneity, and publication bias across studies. Also, we did subgroup analyses to explore the pooled prevalence at different time points and settings. This study was registered with PROSPERO, CRD42021244539. Results. Of 17,115 articles retrieved, 54 studies were included (total sample women). The pooled prevalence of all depression and major depression within one year postpartum was 12.1% (95% CI 10.3%-14.1%; ) and 7.0% (95% CI 5.7%-8.4%; ), respectively. The peaks of all depression occurred during the first 6 months postpartum, especially 2-3 weeks and 6-8 weeks. Subgroup analyses showed that the prevalence of major depression was associated with the income level of countries (higher in low- and middle-income countries (LMICs) than in high-income countries (HICs)) and diagnostic criteria (higher using ICD than using DSM and RDC). No evidence of publication bias was found. Conclusions. Approximately one in eight postpartum women experiences a depressive condition, with one in fifteen suffering major depression. The pooled prevalence based on diagnostic interviews was lower than the existing consensus, which was largely based on self-reported screening instruments. The higher prevalence in LMICs underlines the importance of strengthening research and service provision among these populations.

China recently issued a national plan on perinatal depression (PND) screening. Previous studies elsewhere suggested that uptake of referral after screening for PND is suboptimal, but little is known in China. In this cohort study including 1126 women in Hunan, we identified women at a high risk of PND using the Edinburgh Postpartum Depression Scale (EPDS) over multiple time points. We texted them and offered free consultations with a psychiatrist/psychologist. Among 248 screen-positive women, only three expressed interest and one attended the appointment. We surveyed the women about their reasons for declining referrals and preferred means of care. Of the 161 respondents, 128 (79.5%) indicated that they could cope with the condition without professional assistance and 142 (88.2%) chose their families as the preferred source of help. Only 15 (9.3%) chose professionals as their first option. Implementing a referral policy for screen-positive women would mean approximately one-third of women who gave birth in China would be eligible. Our result argues against referring all screen-positive women for professional services at this time. Interventions should instead build upon the tradition of family support in a more engaged response. These considerations are relevant for the implementation of national screening for PND in China.