Objectives. To examine formal public health educational attainment in state health agency and local health department employees in 2024 and to discuss implications for public health workforce development. Methods. Using data (n = 56 238, unweighted; n = 236 575, weighted) from the 2024 Public Health Workforce Interests and Needs Survey, collected between September 2024 and January 2025, we made inferential comparisons using Rao-Scott adjusted χ2 and fit a logit regression that examined graduate public health degree attainment. Results. Twenty-two percent (12 238/56 238, unweighted; 51 783/236 575, weighted) reported having a public health degree; that proportion is closer to one half among those aged 35 years and younger in the public health sciences. Of public health worker subpopulations, those younger than 50 years old, Asian employees, Black employees, and nonmale staff members (women and other genders) were significantly more likely to have a public health graduate degree than the reference groups. Conclusions. The majority of public health workers do not have public health degrees. Public Health Implications. On-the-job training support should be coupled with deepening academic-practice pathways to ensure the workforce is prepared. (Am J Public Health. Published online ahead of print June 18, 2026:e1-e8. https://doi.org/10.2105/AJPH.2026.308453).
Objectives. To explore rural local health department (LHD) leaders' perspectives on how remote work practices among state health departments (SHDs) may be affecting rural efforts. Methods. For this sequential mixed-methods study, we collected and analyzed qualitative data from 14 rural LHD leaders across the United States in January through April 2025. We also analyzed 2024 Public Health Workforce Interests and Needs Survey data, examining the prevalence of remote work among SHD and LHD staff as well as any significant differences in job satisfaction, retention, and risk of turnover between remote and in-person staff. Results. Participants described impacts of the SHD's remote work options with respect to LHD staff retention and SHD staff's accessibility and familiarity with the LHD's rural needs. Quantitative analyses demonstrated higher rates of remote work among state health staff compared with rural but did not show any significant differences in outcomes of interest, including job satisfaction. Conclusions. Findings point to the ongoing use of and challenges with remote work. Opportunities exist to improve implementation of remote policies and to strengthen support for rural LHDs. (Am J Public Health. Published online ahead of print June 18, 2026:e1-e9. https://doi.org/10.2105/AJPH.2026.308550).
Personal health large language models (PH-LLMs) have rapidly evolved from research prototypes into consumer-facing, data-linked systems that support symptom triage, medication questions, mental health check-ins, and longitudinal self-management. Their direct-to-consumer use without clinical oversight creates a distinct ethical risk profile that general artificial intelligence governance frameworks do not fully address. This viewpoint focuses on text-based, platform-mediated PH-LLMs and synthesizes PH-LLM-specific challenges across 6 domains: privacy, accuracy, equity, transparency, human-artificial intelligence interaction, and regulatory governance. These risks may be amplified by health literacy gaps, longitudinal data aggregation, persuasive conversational design, and fragmented oversight across the consumer-clinical boundary. Grounded in the 4 principles of biomedical ethics, we propose a governance framework that operationalizes beneficence, nonmaleficence, autonomy, and justice through design and deployment controls, including health literacy-aligned communication, crisis and pharmacological safeguards, hallucination mitigation, role disclosure, granular consent, fairness auditing, and accessible design. We further outline implementation mechanisms, including risk-tiered certification, tiered accountability, and postdeployment oversight through adverse-event reporting, transparency reporting, and independent safety evaluation. This framework is intended as an evidence-informed but partly anticipatory approach to governing PH-LLMs in personal health management.
Digital health technologies (DHTs) are increasingly integrated into clinical practice, yet economic evaluations remain scarce, particularly in the early development stages. Within the NICE (National Institute for Health and Care Excellence) Evidence Standards Framework, Tier C DHTs comprise technologies with direct clinical implications and measurable health outcomes, for which robust economic evidence is essential. Early-stage assessments are particularly important to inform subsequent development, refinement, and adoption decisions across the digital health lifecycle. This study aims to explore the feasibility of integrating a full trial-based economic evaluation within an early-stage pilot comparing a chatbot-supported remote patient monitoring (RPM) solution for anticoagulation management with the standard of care (SOC). A cost-effectiveness analysis was performed alongside a pilot crossover trial among adult cardiac surgery patients receiving vitamin K antagonists. Participants were allocated to two 6-month sequences (SOC→RPM or RPM→SOC). The intervention consisted of a rule-based chatbot integrated with home-based international normalized ratio self-testing using portable coagulometers to support communication and therapy management. Effectiveness was measured as time in therapeutic range (TTR), and costs were estimated from the Portuguese National Health Service and a limited societal perspective over a 1-year horizon. The analysis 1 applied a within-patient cost-effectiveness approach to estimate incremental costs, incremental TTR, and incremental cost-effectiveness ratios. Uncertainty was explored through nonparametric bootstrapping (5000 replications) and deterministic sensitivity analyses. Complementary comparisons examined differences between sequences (analysis 2), between periods (analysis 3), and within each sequence (analysis 4). A total of 19 patients were included in the analyses. In analysis 1, RPM improved anticoagulation control, with a mean within-patient increase of 10.43 percentage points in time in TTR. The mean incremental costs were €198.61 (€1=US $1.08) from the Serviço Nacional de Saúde perspective and €270.05 from the limited societal perspective. The corresponding incremental cost-effectiveness ratios were €19.03 and €25.88 per additional percentage point of TTR gained. Sensitivity analyses produced consistent estimates across parameter variations. Complementary analyses (2-4) suggested that RPM tended to be more cost-effective when implemented after the initial 6-month postoperative period. This proof-of-concept study demonstrates that a full trial-based economic evaluation can feasibly be embedded within an early-stage Tier C DHT. The intervention showed improved anticoagulation control alongside higher costs, providing initial insights into its cost-effectiveness profile. Positioned within the digital health evidence continuum, such assessments can function as a learning stage within the lifecycle. To address the persistent adoption-evidence gap, tier- and stage-aligned frameworks are needed to guide the economic evaluation of DHTs. This study contributes to that goal by providing a set of recommendations specifically for Tier C DHTs.
Objectives. To characterize burnout among state and local governmental public health workers and identify implications for public health workforce capacity and sustainability. Methods. We analyzed 2024 Public Health Workforce Interests and Needs Survey (PH WINS) data to assess self-reported burnout and generational differences among US state and local employees. Results. At least 1 symptom of burnout was reported by 71% of respondents. However, burnout was most prominent among Generation Z and Millennials, with 17% and 19% reporting persistent symptoms and 4% and 5% reporting complete burnout, respectively; 10% of Baby Boomers reported persistent symptoms and 1% reported complete burnout. A Rao-Scott adjusted χ2 test showed that the association between burnout and generation was significant, F (7.6, 6674.4) = 89.7, P < .001. Conclusions. High rates of burnout across the PH WINS respondents indicate a systemwide issue. However, higher burnout rates among younger workers highlight the need for targeted interventions and institutionalized changes to support the workforce going forward. Public Health Implications. Systemwide, sustained worker well-being improvements are essential to maintaining a thriving public health workforce, especially as younger generations continue to move into the workforce. (Am J Public Health. Published online ahead of print June 18, 2026:e1-e4. https://doi.org/10.2105/AJPH.2026.308562).
University hospital employees face role-specific stressors that can impair mental well-being and work-related vitality. While worksite health promotion programs show potential for improving mental well-being by targeting lifestyle behaviors, most target single professions or hospital subunits, and evidence for mental well-being and work-related vitality remains mixed. Mobile apps offer unique advantages for delivering such worksite health promotion programs hospital-wide. However, accessible interventions tailored to a diverse workforce are lacking. This study aimed to investigate the feasibility of an app-based worksite health promotion program (the Recharge360 program [The Recharge Company]) targeting multiple lifestyle behaviors, including a team-based competition element, for improving mental well-being and work-related vitality of hospital employees over a 5-month follow-up period by evaluating two objectives: (1) the implementation process of the program, and (2) the preliminary effects of the program on mental well-being and work-related vitality. We included 532 employees (mean age 43, SD 12 y; n=482, 91% women; n=480, 90% highly educated) from a university hospital in Amsterdam, the Netherlands. The study had a single-arm, longitudinal pretest-posttest design lasting 5 months, during which employees participated in the 5-day Recharge360 program (Recharge week) 3 times-in weeks 1, 9, and 17. At baseline (T0) and after each Recharge week (T1-T3), we assessed mental well-being, work ability, need for recovery, and task performance. The process was evaluated by assessing recruitment, attrition, and survey completion rates, and the degree of participation. Preliminary effects were evaluated by linear mixed model regression analyses to assess changes in mental well-being and work-related vitality between baseline and follow-up. Recruitment appeared feasible, but attrition rates were high (up to 70% in the final Recharge week), and the degree of participation decreased over time. We showed statistically significant, albeit small, increases in well-being at T3 (unstandardized β coefficient=2.08, 95% CI 0.33-3.84), with progressively larger improvements in the analyses among those who started at least 1, 2, and all 3 Recharge weeks (unstandardized β coefficient=3.27, 95% CI 1.09-5.45). Results for work-related vitality were mixed. The need for recovery remained unchanged, task performance increased slightly at T3 (unstandardized β coefficient=0.16, 95% CI 0.07-0.24). Work ability showed a small, but statistically significant, decline across follow-up (unstandardized β coefficient=-0.46, 95% CI -0.64 to -0.29). This app-based worksite health promotion program might be feasible to implement in a university hospital setting and shows potential to slightly improve mental well-being, but primarily for a selective group of highly educated, health-conscious women. While these findings support further investigation in a randomized controlled trial in similar university hospital settings, they also highlight the need for more participatory study designs to improve the tailoring of program components and engagement of underrepresented groups, as well as for a supportive culture and population-based approaches at the organizational level.
Drug adherence is crucial for chronic disease management, yet treatment discontinuation remains common due to factors such as side effects, inefficacy, or cost. These reasons are often recorded only in free-text clinical notes, making large-scale analysis difficult. While large language models (LLMs) can interpret such unstructured data more effectively than traditional natural language processing methods, few studies have systematically categorized reasons for discontinuation or identified whether the decision was initiated by the patient or the clinician, especially in low-resource languages such as Estonian. This study aimed to assess the ability of LLMs to extract and classify reasons for drug discontinuation and identify who initiated it using Estonian electronic health records and characterize the observed discontinuation patterns and initiators for statins and antidiabetic medications. We combined prescription data with free-text anamneses from a 10% sample of the Estonian population (2012-2019). LLMs (Llama 3.1-70B and GPT-4o) were applied to extract discontinuation phrases and reasons, classify them into a clinician-developed taxonomy, and identify who discontinued the treatment. Performance was evaluated on 100 randomly chosen cases per drug group. Extraction yielded 625 antidiabetic drug and 233 statin discontinuation cases. Validation confirmed a precision of 0.93 to 0.98 for extracting phrases and 0.95 to 0.96 for extracting reasons. Classification of discontinuation reasons achieved weighted F1-scores of 0.81 to 0.84, whereas classification of who initiated discontinuation achieved weighted F1-scores of 0.64 to 0.78. Adverse reactions were the most frequent reason overall, accounting for 70% (163/233) of statin discontinuations and 44.8% (280/625) of antidiabetic drug discontinuations. Regarding antidiabetic drugs, treatment inefficacy and contraindications were more common. Patients more often stopped due to adverse reactions or nonmedical reasons, whereas physicians more often initiated discontinuation for contraindications. LLMs can accurately extract and classify medication discontinuation reasons and show variable performance in identifying discontinuation initiators in Estonian clinical narratives. Both local and proprietary models showed promising results, enabling scalable analyses that complement structured health records. This demonstrates the potential of LLMs to unlock information from clinical notes, turning this underused electronic health record component into a valuable resource for monitoring treatment patterns and detecting adverse event signals.
Fundus imaging enables noninvasive, high-resolution visualization of the retinal microvasculature. Advances in artificial intelligence (AI) now allow for extraction of quantitative vascular metrics from retinal images, offering new opportunities for identifying systemic health biomarkers. This study sought to characterize retinal microvascular features in a large healthy population and assess their associations with diverse clinical phenotypes and evaluate their ability to predict incident cardiovascular events. We analyzed fundus photographs from 8,467 healthy individuals aged 40-70 years enrolled in the Human Phenotype Project. For external validation we used fundus images from 16,249 participants from UK Biobank. Using an automated AI-based tool (AutoMorph), we extracted 12 quantitative vascular metrics, such as vessel density, average width, fractal dimension, distance tortuosity, and curvature tortuosity, separately for arteries and veins. We derived age- and sex-stratified reference values and evaluated associations with clinical parameters spanning cardiometabolic, respiratory, and behavioral domains. Retinal vascular features demonstrated strong age- and sex-related patterns. Multiple significant associations were observed between microvascular metrics and systemic traits. Arterial features were particularly associated with cardiometabolic factors including blood pressure, lipid profiles, glycemic indices, and body composition (body mass index, fat mass), as well as sleep apnea parameters. Findings replicated in UK Biobank and demonstrated prognostic value for incident cardiovascular events. This large-scale, AI-driven study provides normative data on retinal vascular traits and supports the utility of fundus imaging for systemic risk stratification and prediction of cardiovascular events. Our findings highlight the potential of retinal biomarkers for early detection and monitoring of cardiometabolic and sleep-related conditions, reinforcing the emerging role of oculomics in predictive and preventive health care.
To be beneficial for empirical health research, a dataset must be fit for use. The quality of a dataset can only be influenced during data collection, yet it is evaluated multiple times during analysis or secondary use by applying quality indicators. This study aimed to establish an up-to-date set of indicators measuring the quality of datasets in empirical health research. A total of 3 pillars were combined. First, the 51 indicators of a German guideline from 2014 about the management of data quality were revised. Second, a literature review was performed looking for evidence sources since 2013 that describe, propose, or apply dataset quality indicators. Third, indicators were supplemented by a manual search and other sources. The quality indicators were then integrated into the IDEFIM framework. The IDEFIM framework distinguishes between the categories' data, metadata, context, and openness quality. In this work, only the categories data and metadata quality, with their 14 dimensions were considered. In total, 69 indicators qualified for the IDEFIM indicator set, 53 related to the category data quality, and 16 to the category metadata quality. A total of 30 indicators originated from the German guideline, 31 from the literature review. Three indicators were added to cover aspects of diversity, equity, and inclusion, and an additional 5 related to specifics of data and metadata quality not addressed so far. Most indicators were found in the dimensions accuracy (data) with 12 measures, completeness (data) with 12 measures, and consistency (data) with 19 measures. According to the number of supporting evidence sources, missing values in data elements (48 evidence sources), contradictions (31), and currentness (26) were the most popular quality indicators. Metadata quality was significantly less frequently addressed. The presented IDEFIM indicator set can be used for the management of data collections as well as for the verification of a dataset's quality for an intended use. The indicator set should also be considered in the design of a study in empirical health research and the development of software tools supporting the visualization of issues related to the quality of a dataset.
An increasing amount of digital health data are being collected across rehabilitation settings, but their integration into routine clinical practice remains limited, despite its potential to motivate patients or inform clinical decision-making. Specifically, effective visualization and communication of assessment outcomes to both patients and health care practitioners (HCPs) represent a key gap in the neurorehabilitation practice. This study describes the development and evaluation of RehaLink (author ND, ETH Zürich), a proof-of-concept mobile app that delivers structured, interpretable feedback from conventional and technology-based assessments to neurorehabilitation patients and HCPs. The app was developed through a 3-step iterative co-design process involving 17 inpatients with multiple sclerosis and 15 HCPs from a single rehabilitation center. The app integrates a full battery of conventional assessments routinely conducted at the clinic, as well as digital health metrics from the Virtual Peg Insertion Test, a validated technology-based assessment of upper limb function, as a proof of concept for integrating technology-based assessment data into clinical workflows. Three structured feedback sessions were conducted, in which participants evaluated feedback types, visualization formats, and app usability using Likert-scale ratings, preference rankings, open-ended questions, and the System Usability Scale. Data were analyzed using descriptive statistics and directed content analysis. Across all 3 sessions, progress bars and color-coded indicators were consistently preferred over text-heavy or abstract formats by both patients and HCPs. A persistent set of competing demands was observed, with participants requesting both visual simplicity and access to absolute values and normative comparisons. HCPs tended to underestimate patients' preference for informative visualizations. The perceived value of structured feedback increased over the course of the study; patients' median ratings rose from 4.0 to 5.0 and HCPs' from 4.0 to 4.5 on a 5-point Likert scale. The resulting mobile app prototype demonstrated high usability, with patients achieving a mean System Usability Scale score of 93.6 (mean 6.4; best imaginable) and HCPs 80.9 (SD 8.1; good), according to established benchmarks. These findings demonstrate the feasibility and value of a co-designed digital feedback tool for neurorehabilitation. By combining conventional and technology-based assessment outcomes in an accessible, user-centered format, the app has the potential to enhance patient engagement, support clinical decision-making, and advance the implementation of value-based, personalized care.
A 36-item short-form (SF-36) was constructed to survey health status in the Medical Outcomes Study. The SF-36 was designed for use in clinical practice and research, health policy evaluations, and general population surveys. The SF-36 includes one multi-item scale that assesses eight health concepts: 1) limitations in physical activities because of health problems; 2) limitations in social activities because of physical or emotional problems; 3) limitations in usual role activities because of physical health problems; 4) bodily pain; 5) general mental health (psychological distress and well-being); 6) limitations in usual role activities because of emotional problems; 7) vitality (energy and fatigue); and 8) general health perceptions. The survey was constructed for self-administration by persons 14 years of age and older, and for administration by a trained interviewer in person or by telephone. The history of the development of the SF-36, the origin of specific items, and the logic underlying their selection are summarized. The content and features of the SF-36 are compared with the 20-item Medical Outcomes Study short-form.
暂无摘要（点击查看详情）
Preoperative cardiovascular risk stratification is essential in noncardiac surgery, but conventional testing is frequently overused, increasing costs without improving outcomes. Artificial intelligence (AI)-enabled electrocardiography (ECG) may enhance perioperative risk assessment by identifying surgical candidates at very low-risk for adverse events. This study aimed to evaluate whether AI-ECG-based risk stratification could help decrease low-yield preoperative cardiovascular testing and reduce associated costs, without an observed increase in postoperative adverse outcomes, in candidates for noncardiac surgery. We retrospectively analyzed 41,218 patients (46,135 ECG-surgery pairs) undergoing noncardiac surgery at Seoul National University Bundang Hospital (2020-2021). An AI-ECG algorithm generated eight probability scores for cardiac conditions, classifying cases as low- or high-risk. Based on the performance and results of preoperative cardiovascular testing (transthoracic echocardiography, coronary computed tomography angiography, single-photon emission computed tomography, or coronary angiography), cases were classified as no advanced cardiovascular imaging, negative-test, or positive-test. The primary end point was a 30-day composite of all-cause mortality and unplanned percutaneous coronary intervention. AI-ECG classified 92.3% (42,599/46,135) of cases as low-risk, with a composite outcome rate of 0.2% (79/42,599) vs 2.9% (101/3536) in high-risk cases. Preoperative cardiovascular testing was performed in 11.8% (5458/46,135) of cases, with only 16.3% (892/5458) yielding positive findings. In AI-ECG low-risk cases, event rates were uniformly low (0.2%-0.4%) irrespective of whether advanced cardiovascular testing was performed, whereas in high-risk cases, rates were consistently high (2.6%-3.4%). The incidence of the composite outcome was consistently higher in AI-ECG-graded high-risk cases across all European Society of Cardiology surgical risk and Revised Cardiac Risk Index strata. In this retrospective cohort, a multitask AI-ECG identified surgical candidates at low-risk for postoperative complications, for whom advanced cardiovascular testing demonstrated low diagnostic yield. Integrating AI-ECG with conventional risk tools may offer an exploratory strategy to optimize resource use and minimize redundant testing. Prospective studies are needed to confirm the clinical and economic benefits of AI-ECG as a screening tool.
暂无摘要（点击查看详情）
暂无摘要（点击查看详情）
暂无摘要（点击查看详情）
The mandatory collection of patient-reported outcome measures and the implementation of thresholds for hip and knee replacement surgery represent a growing international trend in value-based health-care policy. Our aim was to investigate whether the Oxford Hip Score (OHS) and Oxford Knee Score (OKS) can be used to accurately predict patient satisfaction, and to estimate thresholds to guide value-based health-care policy. All primary total hip replacements (THRs) and total knee replacements (TKRs) for osteoarthritis undertaken at a tertiary academic institution over a 6-year period were identified. Logistic regression models were used to evaluate preoperative and postoperative values for the OHS and OKS, and the change between them, as predictors of patient satisfaction. Optimal thresholds for both the minimal clinically important difference (MCID) and the substantial clinical benefit (SCB) were identified. A total of 1,429 THRs (mean patient age, 66.1 years, standard deviation [SD], 11.1 years; 819 [57.3%] female) and 1,079 TKRs (mean patient age, 68.3 years, SD, 8.4 years; 635 [58.9%] female) were included. For the postoperative OHS, thresholds of 35.5 (95% confidence interval [CI], 29.1 to 41.9) for the MCID and 36.5 (95% CI, 33.0 to 40.0) for the SCB were identified. For the postoperative OKS, thresholds of 30.5 (95% CI, 24.2 to 36.8) for the MCID and 38.5 (95% CI, 36.7 to 40.3) for the SCB were identified. For the change in OHS, thresholds of 19.5 (95% CI, 13.4 to 25.6) for the MCID and 20.5 (95% CI, 16.0 to 25.0) for the SCB were identified. For the change in OKS, thresholds of 13.5 (95% CI, 8.3 to 18.7) for the MCID and 14.5 (95% CI, 11.7 to 17.3) for the SCB were identified. Patients with worse preoperative function had higher thresholds. Preoperative Oxford scores were poor predictors of patient satisfaction. The thresholds for the postoperative Oxford scores and change scores may guide value-based health-care decision-making using the OHS and OKS. Prognostic Level II. See Instructions for Authors for a complete description of levels of evidence.
Loneliness has been linked to reduced mental and physical health. The "loneliness epidemic" is recognized as a public health crisis. However, questions remain about the potential of video games, which people play by themselves, to help reduce perceived loneliness. This study explored the extent to which open-world games (eg, The Legend of Zelda: Breath of the Wild) and fun, accessible games (eg, Yoshi's Crafted World) can help reduce loneliness in adults. We examined how such gameplay can foster a stoic approach to life and how stoicism mediates the reduction of perceived loneliness. A cross-sectional survey was conducted using convenience sampling near a video game store. The sample comprised 2252 adults aged 21 years and older (women: n=966, 42.90%; men: n=1281, 56.90%; prefer not to disclose: n=5, 0.20%). Participants completed a questionnaire to measure perceived loneliness and stoicism, as well as their gameplay habits. Data were analyzed using ANOVA and moderated mediation with the PROCESS macro (bootstrapped samples=5000; 95% CI) to examine the effects of video gameplay on stoicism and loneliness, with the α level set at .05. Zelda players indicated higher stoicism (mean 4.87, SD 0.11; 95% CI 4.66-5.08) than nonplayers (mean 3.23, SD 0.07; 95% CI 3.09-3.37; F1,2252=164.95; P<.001). Yoshi players also noted significantly higher stoicism (mean 4.49, SD 0.12; 95% CI 4.27-4.72) than nonplayers (mean 3.61, SD 0.05; 95% CI 3.50-3.71; F1, 2252=48.33; P<.001), with a significant interaction effect (F1,2252=7.89; P=.005) on stoicism. Furthermore, Zelda players indicated lower loneliness (mean 3.02, SD 0.11; 95% CI 2.81-3.22) than nonplayers (mean 4.28, SD 0.07; 95% CI 4.14-4.42; F1, 2252=98.32; P<.001). Yoshi players also noted significantly lower loneliness (mean 3.09, SD 0.12; 95% CI 2.86-3.32) than nonplayers (mean 4.21, SD 0.05; 95% CI 4.10-4.32; F1, 2252=76.32; P<.001). Moderated mediation analysis demonstrated that Zelda gameplay was positively associated with stoicism (β=1.28, 95% CI 1.07-1.50; P<.001), and stoicism was negatively associated with perceived loneliness (β=-0.49, 95% CI -0.52 to -0.45; P<.001). This study is innovative in identifying stoicism as a potential emotional pathway through which video games may reduce loneliness. Moving beyond views of gaming as passive escapism, our findings suggest that specific gameplay experiences may serve as active spaces for cultivating resilience. We introduce a "digital diet" framework, indicating that balancing open-world challenges (eg, Zelda) with low-stakes restoration (Yoshi) produces synergistic psychological support. Practically, thoughtfully curated gaming experiences may serve as scalable and cost-effective digital adjuncts for public mental health interventions addressing the loneliness epidemic.
We adopted a life course perspective to characterize lifetime marital history typologies and examined their associations with later-life cognitive function in two nationally representative cohorts of adults aged 65+ in the United States and China. We also explored whether the associations vary by gender. We used data from the Harmonized Cognitive Assessment Protocol from the Health and Retirement Study (HRS-HCAP) (n=3,318) and the China Health and Retirement Longitudinal Study (CHARLS-HCAP) (n=5,766). Cognition was measured by general cognitive function, executive function, memory, language, and orientation. Country-specific marital histories from age 18 to 65 were constructed using sequence and cluster analysis. Population-weighted linear regression models examined the associations of marital histories with general and domain-specific cognitive outcomes. Marital histories in the United States are more heterogeneous than in China. Relative to lifelong marriage with a normative age at onset (mid- to late-20s and 30s in both the United States and China), lifelong marriage with an onset in early 20s and premature widowhood by age 50 were associated with worse cognitive function in both countries. American older adults divorced and remarried, or divorced by age 40 without remarrying, did not perform worse across all cognitive outcomes, compared to the normatively married. Lifelong singlehood was associated with worse general cognitive function in China, but not in the United States. Our findings demonstrate that marital histories, including timing, sequence, and transition, have enduring implications for later-life cognitive health, with patterns shaped by cultural norms and the timing of marital transition.
While cross-sectional studies have consistently reported an association between nonsuicidal self-injury (NSSI) and internet addiction (IA), longitudinal evidence regarding the directionality and dose-response relationship remains limited. Furthermore, the roles of sex and varying degrees of problematic internet use in predicting new-onset NSSI are not fully understood. This prospective cohort study aimed to investigate whether baseline IA and its intermediate states predict the subsequent new onset of NSSI among Chinese adolescents over a 6-month period and to explore potential sex differences in this longitudinal association. A prospective cohort design was used. A total of 1315 junior high school students without a history of NSSI were recruited at baseline, and 704 (53.5%) students completed the 6-month follow-up. IA and NSSI were assessed using the Chinese Internet Addiction Scale-Revised and a self-report questionnaire from the Adolescent Health-Related Risky Behavior Inventory, respectively. Logistic regression analysis was conducted to examine the predictive value of IA exposure for incident NSSI, adjusting for key covariates, including sex, age, ethnicity, only child status, anxiety, and depression. Restricted cubic spline regression was used to model the dose-response relationship between distinct states of IA and NSSI risk. The baseline prevalence of IA was 9.09% (64/704). At the 6-month follow-up, the incidence rate of NSSI was 9.8% (69/704). Restricted cubic spline regression revealed a linear dose-response relationship, where the risk of incident NSSI escalated with increasing IA severity. In the fully adjusted model for the total sample, baseline IA was a significant predictor of subsequent NSSI (odds ratio [OR] 2.185, 95% CI 1.031-4.627; P=.04). Crucially, stratified analyses revealed significant sex disparities: the longitudinal association between IA and subsequent NSSI was statistically significant among female adolescents (OR 3.271, 95% CI 1.101-9.717; P=.03) and the intermediate internet-dependent state (OR 2.593, 95% CI 1.002-6.710; P=.049) but not among male adolescents (IA: P=.44; internet-dependent state: P=.87). NSSI incidence is notably prevalent among Chinese junior high school students. While IA serves as a robust, independent risk factor for predicting the new onset of NSSI in the overall adolescent population, sex-stratified analyses revealed that this longitudinal association was statistically significant (P=.03) only among female students. These findings underscore the critical need to integrate IA assessments into school-based mental health screenings and highlight the necessity of developing sex-specific, emotion-focused prevention strategies to mitigate NSSI risk.