Empirical research so far largely suggests that population policies around the world do not explain a majority of the global fertility decline. Notably, even China’s total fertility rate declined little after implementation of the one-child policy in 1979 and 1980, but fertility then fell sharply by more than one-third during the early 1990s. Here we show that strengthening bureaucratic incentives under China’s one vote veto policy, which strictly prohibited career advancement for adherence failure, was instrumental in more effective implementation of the one-child policy, and in its delayed impact on fertility. Using provincial variation in implementation timing, we find that the one vote veto explains nearly half of China’s total fertility rate decline during the 1990s, driving fertility below the replacement rate. Increases in use of intrauterine devices (the most prevalent form of modern contraception in China) that were recommended by local officials were particularly salient. Overall, our findings suggest that population policy, when more strictly enforced due to stronger bureaucratic incentives, made a greater contribution to low fertility in China than suggested by past research.
Preprimary education is central to the global education agenda, including Sustainable Development Goal 4.2. Despite global momentum in low- and middle-income countries, cross-national evidence on long-term outcomes associated with preprimary education legislation remains limited. Here we applied a matrix-completion approach with subgroup and sensitivity analyses using 495 nationally representative surveys from 1990 to 2024, covering nearly 6.7 million women across 117 countries. Eligibility under the legislation was associated with higher educational enrolment and completion, and with more gender-egalitarian attitudes, more favourable empowerment-related behaviours and greater reproductive autonomy. For example, eligibility was associated with higher female primary-school enrolment (1.21 percentage points (pp); 95% confidence interval (CI) 0.52 to 1.91; P < 0.001) and completion (0.75 pp; 95% CI 0.24 to 1.25; P = 0.004), greater disapproval of wife-beating (1.04 pp; 95% CI 0.53 to 1.56; P < 0.001), greater participation in decisions about husbands’ earnings (3.25 pp; 95% CI 2.50 to 4.00; P < 0.001) and greater ever use of contraception (2.96 pp; 95% CI 2.33 to 3.59; P < 0.001). Eligibility under the legislation was also associated with lower rates of adolescent childbearing (−2.41 pp; 95% CI −2.67 to −2.14; P < 0.001) and child marriage (−0.43 pp; 95% CI −0.71 to −0.15; P = 0.003). These findings provide cross-national evidence directly relevant to ongoing policy discussions surrounding Sustainable Development Goal 4.2 and broader global development agendas, while emphasizing the importance of implementation capacity and equitable system expansion.
Objectives We examined the effectiveness of a school-based intervention focusing on sugar-sweetened beverage (SSB) harms and nutrition-label understanding among children and adolescents.
Study design We conducted a school-based, two-arm, cluster-randomised controlled trial among upper-primary and lower-secondary students in six Chinese schools. The trial was embedded within the 2022 wave of the Health and Nutrition Panel Survey (HNPS-2022). Classes were the unit of randomisation; within each school and target grade, classes were randomly assigned to a video (<5 min) on SSB harms and nutrition labels or eyesight protection. The trial included a real choice experiment (RCE) and a discrete choice experiment (DCE). Primary outcomes were the probability of selecting a full-sugar fermented milk drink (FMD) at actual beverage selection in the RCE, and the direction and relative importance of preferences for prespecified beverage attributes in the DCE. We used ANCOVA and mixed logit models.
Results Between September and October 2022, 24 classes were randomly assigned to control (12 classes; 653 students) or intervention (12; 649); 1269 students were analysed (641 control and 628 intervention). At actual beverage selection, 367/641 (57.3%) control participants selected the full-sugar rather than the low-sugar FMD, compared with 228/628 (36.3%) intervention participants (adjusted absolute difference −0.21, 95% CI -0.28 to −0.15; p < 0.0001; adjusted odds ratio 0.39, 95% CI 0.30 to 0.51; p < 0.0001). In the DCE, beverage profile (a combination of beverage type and sugar content) had the highest relative importance (49.6%), followed by price (38.4%). The intervention shifted preferences towards healthier options, especially among primary school students.
Conclusions A brief health literacy intervention supported healthier immediate beverage choices, but sustained SSB prevention will require clearer, child-friendly sugar information and school policies that make healthier beverage options easier to choose.
Background The Thinking Healthy Extended Program (THEP) is a nonspecialist-delivered mental health intervention targeting female caregivers (mothers and grandmothers) of children aged 6-36 months in rural China. It consists of 24 fortnightly group sessions, supplemented by individual home visits for caregivers who miss group sessions, and integrates cognitive reframing techniques, behavioral engagement, childrearing skills, and peer support. Embedded within a large-scale randomized controlled trial, this study aims to evaluate implementation fidelity of THEP, examine hypothesized mechanisms of change, and identify contextual factors, including enablers and barriers, that influenced intervention implementation.
Methods In-depth interviews were conducted at 3 months and 10 months into the intervention implementation with 50 caregivers and 10 facilitators. The interview guides were developed based on the UK Medical Research Council’s process evaluation framework, with the facilitator interview guide focused on implementation fidelity, the caregiver interview guide focused on mechanisms of change, and contextual factors explored in both guides. All interviews were audio-recorded, transcribed, and analyzed using thematic analysis.
Results THEP facilitators demonstrated competence in coordinating schedules, delivering sessions, and engaging caregivers. Caregivers reported improvements in peer support and childrearing skills, along with some uptake of behavioral engagement. However, both facilitators and caregivers demonstrated limited understanding of cognitive reframing techniques. Grandmothers (who accounted for 43% of the sample in the current study) experienced greater challenges than mothers with respect to applying the four program mechanisms of change. Several contextual factors, including a child-focused orientation, distractions related to the presence of children at sessions, and intergenerational differences between mother and grandmothers, influenced THEP implementation.
Conclusions THEP was feasible to deliver through nonspecialist facilitators, but implementation quality was constrained by limited understanding of cognitive reframing strategies and variation in caregiver engagement. Future adaptation and scale-up should prioritize facilitator capacity and intervention approaches that are more feasible for nonspecialist providers and more accessible to diverse caregivers.
The emergence of artificial intelligence (AI) has heightened interest in personalizing computer assisted learning (CAL) programs to tailor their instruction to individual students. Despite the proliferation of AI-driven CAL programs, evidence for their effectiveness remains limited. We present findings from a field experiment in rural China examining whether an AI-driven personalized CAL program improves student achievement. We randomly assign 8647 students from 315 primary school classes to one of three treatment arms: (i) AI-CAL, (ii) non-personalized CAL (active control), and (iii) non-CAL educational activities (pure control). Results indicate that AI-CAL does not significantly improve achievement, with estimates precise enough to rule out non-trivial positive effects. This finding holds across the difficulty of assessment items and across the baseline achievement of students. If anything, AI-CAL might be less effective than non-personalized CAL for students in the middle tercile of baseline achievement. These findings call for caution when scaling AI-driven educational technologies.
International Journal for Equity in Health,
October 14, 2026
Background: Early motherhood is a critical yet vulnerable life stage shaped by intersecting individual, familial, societal and structural challenges. Existing research often overlooks the multiple everyday stressors mothers experience across ecological domains. This study explores the lived experiences of mothers in urban China, using an ecological framework to examine the multi-level challenges they face during the first three years after childbirth.
Methods: In-depth interviews were conducted with 50 mothers of children aged 0–3 years recruited from nine community and township health centers and one childcare program for a migrant community in Shanghai, China, during well-child visits. Interviews were guided by an ecological framework and analyzed using thematic analysis.
Results: Challenges clustered across ecological levels, including postpartum physical health problems, caregiving burden, limited family and social support, work–childcare tension and constrained services, and information and financial constraints. These challenges manifested as common themes that negatively affected maternal wellbeing and the child nurturing environment.
Conclusions: These findings highlight that the challenges mothers experience extend beyond the individual level to family and institutional contexts, underscoring the need for coordinated interventions from individual support to family- and system-level services.
This study examines whether and how caregiver mental health symptoms influence the impacts of a center-based parenting intervention on child development.
METHODS
We conducted a clustered randomized controlled trial across 100 rural villages in China between 2016 and 2019. Villages were randomly assigned to a treatment group that received weekly 1-on-1 parenting sessions through parenting centers or to a pure comparison group. A total of 2554 caregiver–child dyads with children aged 6 to 24 months were surveyed at baseline, and 2044 completed the endline survey after 1 year (follow-up rate: 80.0%). Child development was assessed using the Bayley Scales of Infant Development, Third Edition. Caregiver mental health was measured using the Depression Anxiety and Stress Scale-21. Ordinary least squares regressions with interaction terms were used for analysis.
RESULTS
The intervention significantly improved cognitive (β = 0.17 SD; 95% CI 0.045–0.29) and language development (β = 0.10 SD; 95% CI 0.00–0.21) among children whose caregivers did not exhibit depressive symptoms at baseline. Positive cognitive effects were also observed among children of caregivers with low anxiety (β = 0.13 SD; 95% CI 0.00–0.26) or low stress (β = 0.15 SD; 95% CI 0.03–0.27). In contrast, no significant effects were detected across developmental domains among children whose caregivers exhibited mental health symptoms. Heterogeneous impacts were driven by differences in stimulating parenting practices rather than program attendance or parenting beliefs.
CONCLUSIONS The intervention primarily benefited children whose caregivers had no baseline mental health symptoms, suggesting caregiver mental health moderates intervention effects.
Generative artificial intelligence (AI) systems now summarize health-related search results, answer medical questions and offer guidance people once sought from clinicians. These systems bring real benefits, including plain-language explanations of medical information, around-the-clock availability and expanded access for people facing language or literacy barriers. They also carry new risks: inaccurate guidance can harm people at scale, and malicious actors can now generate personalized health misinformation at negligible cost. In this Perspective, we argue that these risks are largely invisible to the institutions responsible for protecting public health. When AI guidance causes harm, no record exists outside the platform, no channel allows users to report it and no independent researcher can measure the consequences. We trace these invisible risks across two settings: incidental exposure online and active seeking through search engines and chatbots. Minimizing harm from AI-generated health information requires making it observable. We therefore offer recommendations that aim to improve transparency, mitigate harm at the point of delivery and assign accountability, ranging from voluntary platform measures to regulatory ones.