Tag: pubmed

BMC Public Health. 2026 Feb 5. doi: 10.1186/s12889-026-26453-x. Online ahead of print.

ABSTRACT

BACKGROUND: Despite the crucial distinction between insomnia symptoms and a diagnosed disorder, population-level studies based on contemporary criteria and clinical interviews are scarce. This study therefore examined the insomnia spectrum by assessing the prevalence, identifying subtypes, and exploring associations with sociodemographic factors and comorbid mental disorders for both conditions.

METHODS: This large-scale, community-based cross-sectional study was conducted in Beijing from October to December 2021. A sample of 10,778 adults was recruited via multistage stratified random sampling. Trained psychiatrists conducted standardized diagnostic interviews based on the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) to collect data on insomnia disorder and mental disorders. Descriptive analysis and weighted Rao-Scott chi-square tests, were performed using the Statistical Analysis System (SAS, version 9.4).

RESULTS: The weighted prevalence rates were 17.7% for subjective sleep problems, 10.7% for clinically assessed insomnia symptoms, and 3.0% for DSM-5-diagnosed insomnia disorder. Among insomnia subtypes, initial insomnia was most prevalent, both as a symptom (8.1%) and a disorder (2.6%). Furthermore, insomnia disorder prevalence varied by sociodemographics, being higher in females, older adults (≥ 60 years), and those with lower education. Among individuals with insomnia disorder, 31.3% had comorbid other mental disorders, particularly alcohol-related disorders (13.4%). Conversely, insomnia disorder was observed in 14.4% of individuals with other mental disorders, with the highest prevalence in depressive (21.4%) and anxiety disorders (19.2%).

CONCLUSIONS: The marked disparity between prevalent insomnia symptoms and formal diagnoses, compounded by significant psychiatric comorbidity, mandates a public health shift toward population-level screening and early intervention. This imperative includes adopting standardized diagnostics and implementing integrated, transdiagnostic treatment models as a pivotal preventive strategy.

PMID:41639643 | DOI:10.1186/s12889-026-26453-x

BMC Geriatr. 2026 Feb 4. doi: 10.1186/s12877-026-07011-x. Online ahead of print.

ABSTRACT

INTRODUCTION: Older adults are more likely than younger people to have multiple chronic health conditions and increased health and/or social needs. As older people generally prefer living at home in the community as they age and residential care can be expensive, there is a need for effective alternatives to residential care in the community. The objective of this review was to synthesize evidence about programs aimed at enabling older people with ongoing health and social care needs to remain in the community.

METHODS: This review followed the JBI methodology for systematic reviews of effectiveness. Included studies reported on complex, multifactorial interventions that were based in the community and included more than one type of service. Six databases and gray literature were searched for published and unpublished research. Titles and abstracts, and full-text selections were screened by two or more reviewers and assessed for methodological quality using JBI critical appraisal tools. Results related to quality of life and healthcare outcomes were extracted.

RESULTS: Fifty-five full text articles, reporting on 51 unique complex interventions, were included in the review. Studies were predominantly randomized controlled trials (n=24) and quasi-experimental studies (n=23), with five cohort and three case series studies included. The overall quality of the included studies was moderate. Key characteristics of the interventions included case management, care planning, a comprehensive assessment, and in-home visits. Comparative meta-analyses were completed for five of the outcomes (hospital admission, emergency department visits, long-term care use, primary care use and quality of life). The results showed effects in the direction of interventions for the number of hospital admissions and LTC use, however, none of the meta-analyses were statistically significant.

CONCLUSIONS: There is little agreement about the effectiveness of complex interventions on quality of life and health system outcomes. Jurisdictional differences may make the integration of literature reporting on such interventions particularly difficult. There is an ongoing need to understand what helps older people with complex needs live well in the community and what level of health system engagement is optimal.

SYSTEMATIC REVIEW REGISTRATION: PROSPERO reference number CRD42022324061.

PMID:41639636 | DOI:10.1186/s12877-026-07011-x

BMC Med Res Methodol. 2026 Feb 4. doi: 10.1186/s12874-025-02762-4. Online ahead of print.

ABSTRACT

BACKGROUND: Analyses of craniofacial morphology are essential for various medical and research applications, including the study of midfacial development, dysmorphologies, and planning surgical interventions. Incomplete CT scans often due to patient movement, imaging artifacts, or obscured landmarks which can result in missing data. If not properly addressed, such missingness may bias conclusions and weaken statistical power.

OBJECTIVE: This paper evaluates imputation techniques to identify the most suitable method for handling missing completely at random values in small, high-dimensional, and highly correlated craniofacial morphometric datasets.

METHODS: 42 craniofacial variables were measured from 32 observations. The missing data structure was set to be at random with 268 (20%) missing values. Five common imputation techniques namely Mean/Median imputation, k-Nearest Neighbors (kNN), Multiple Imputation by Chained Equations (MICE), Random Forest (RF), and Decision Tree, were considered. The performance of the imputation technique was quantified using Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Variance Preservation.

RESULTS: RF Imputation demonstrated the best overall performance, with the lowest RMSE (1.3987) and MAE (0.4902), indicating a high level of accuracy in imputing missing values. It also maintained a relatively close to 1 variance preservation (0.8961), suggesting its effectiveness in retaining the original variability in the dataset. MICE present lower accuracy with high RMSE (3.0869) and MAE (1.1246) however appear to have the closest variance preservation to 1 (1.0580).

CONCLUSION: The findings emphasize the importance of choosing suitable imputation techniques for small, high-dimensional, and correlated datasets such as those in craniofacial morphometry. RF emerged as the most effective method, offering a strong balance between accuracy and variance preservation.

PMID:41639629 | DOI:10.1186/s12874-025-02762-4

Nature. 2026 Feb;650(8100):110-115. doi: 10.1038/s41586-025-09951-7. Epub 2026 Feb 4.

ABSTRACT

Fast and reliable validation of new designs in complex physical systems such as batteries is critical to accelerating technological innovation. However, battery development remains bottlenecked by the high time and energy costs required to evaluate the lifetime of new designs^1,2. Notably, existing lifetime forecasting approaches require datasets containing battery lifetime labels for target designs to improve accuracy and cannot make reliable predictions before prototyping, thus limiting rapid feedback^3,4. Here we introduce Discovery Learning, a scientific machine learning approach that integrates active learning⁵, physics-guided learning⁶ and zero-shot learning⁷ into a human-like reasoning loop, drawing inspiration from educational psychology. Discovery Learning can learn from historical battery designs and reduce the need for prototyping, thereby predicting the lifetime of new designs from minimal experiments. To test Discovery Learning, we present industrial-grade battery data comprising 123 large-format lithium-ion pouch cells, including diverse material-design combinations and cycling protocols. Trained on public datasets of cell designs different from ours, Discovery Learning achieves 7.2% test error in predicting cycle life using physical features from the first 50 cycles of 51% of cell prototypes. Under conservative assumptions, this results in savings of 98% in time and 95% in energy compared with conventional practices. Discovery Learning represents a key advance in accurate and efficient battery lifetime prediction and, more broadly, helps realize the promise of machine learning to accelerate scientific discovery⁸.

PMID:41639573 | DOI:10.1038/s41586-025-09951-7

Subst Use Misuse. 2026 Feb 4:1-8. doi: 10.1080/10826084.2026.2621260. Online ahead of print.

ABSTRACT

BACKGROUND: This study examined illicit substance use among youth in Taiwanese temple parade troupes, a culturally distinct but high-risk population.

METHODS: In 2019, we surveyed 696 participants recruited on-site through convenience sampling across Taiwan using a structured questionnaire that assessed demographics, alcohol and tobacco use, substance-involved social ties, and acquisition settings. Three domains of substance-involved ties (family, peers, intimate partners; range 0-3) and three types of acquisition settings (private, social, public; range 0-3) were analyzed using descriptive statistics, one-way ANOVAs, and logistic regression models. The lifetime prevalence of illicit substance use was 9.8%. Because only three female participants reported illicit substance use, multivariable logistic regression analyses were restricted to male participants (n = 593) and adjusted for age and education.

RESULTS: Odds of illicit substance use increased with the accumulation of substance-involved social relationships, with adjusted odds ratios (aORs, 95% CI) of 2.96 (1.48-5.91) for one type and 11.32 (3.72-34.44) for two types, compared with none; estimates for three types could not be obtained due to complete separation. In contrast, awareness of any acquisition setting was associated with markedly higher odds of illicit substance use (aORs, 95% CI: 1 = 22.48, 5.25-96.34; 2 = 59, 3.78-81.79; 3 = 37.27, 4.75-292.23).

CONCLUSIONS: Findings highlight that accumulated substance-involved relationships and access through any acquisition setting are key structural conditions shaping substance use in this cultural context. While limited by its cross-sectional design and convenience sampling, the study provides novel evidence on culture-embedded risk structures and underscores the need for longitudinal and context-sensitive prevention strategies.

PMID:41637117 | DOI:10.1080/10826084.2026.2621260

JAMA Dermatol. 2026 Feb 4. doi: 10.1001/jamadermatol.2025.5723. Online ahead of print.

ABSTRACT

IMPORTANCE: Recessive dystrophic epidermolysis bullosa (RDEB) is a rare monogenic blistering disorder with wide clinical heterogeneity, ranging from localized skin fragility to life-limiting systemic complications. Understanding genotype-phenotype correlations in COL7A1, the causative gene, is critical for clinical prognostication, genetic counseling, and the rational design of emerging molecular therapies.

OBJECTIVE: To determine the frequency of genotypic and phenotypic subtypes, and to assess whether variant type or location can predict phenotypic severity and extracutaneous complications in patients with RDEB carrying homozygous variants.

EVIDENCE REVIEW: This was a systematic review of all RDEB genotypes and phenotypes reported to the International Dystrophic Epidermolysis Bullosa Patient Registry (DEB Registry) and eligible studies published in English from May 1993 to September 2025. PubMed, Cochrane Library, and Web of Science were searched and eligible studies were reviewed following PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020) guidelines. Included studies reported bi-allelic COL7A1 variants and clinical phenotypes. Data from the DEB Registry were cross-checked to supplement the published cases. Descriptive statistics were used for data analyses, and Fisher exact and χ2 methods were used to test additional genotype-phenotype correlations in patients with RDEB carrying homozygous variants.

FINDINGS: A total of 1802 patients with RDEB comprising 1002 pathogenic variants within COL7A1 were identified from 217 articles. Among the 706 patients with homozygous variants (mean [SD; range] age, 12.2 [13.0; 0-72] years), 533 (75.5%) had severe RDEB, most frequently associated with frameshift and nonsense variants (388 [72.8%] premature termination codons [PTCs]). In contrast, intermediate and milder subtypes were associated with missense or non-PTC variants. Variant location also influenced phenotype: homozygous variants affecting the noncollagenous 1 domain were associated with severe RDEB in 74 of 83 unique variants (89.2%). Extracutaneous involvement clustered in homozygous PTC carriers and was observed almost exclusively in severe RDEB, with occasional cases in the intermediate subtype and rare instances in the inversa, localized, and self-improving subtypes. Recurrent and population-specific variants suggested founder effects. Splice site and missense variants showed phenotypic variability, with augmented intelligence-based predictions correlating with severity.

CONCLUSIONS AND RELEVANCE: In this systematic review, the type and site of pathogenic variants in COL7A1 correlated with the severity of RDEB phenotype across different nationalities, races, and ethnicities. These findings may provide improved patient prognosis, genetic counseling, and personalized therapeutics.

PMID:41637086 | DOI:10.1001/jamadermatol.2025.5723

Acta Odontol Scand. 2026 Feb 4;85:61-66. doi: 10.2340/aos.v85.45419.

ABSTRACT

OBJECTIVE: This study aimed to evaluate the awareness levels of actively practicing dentists in Türkiye regarding artificial intelligence (AI)-related ethical issues, data security, anonymization, and legal regulations.

MATERIALS AND METHODS: A cross-sectional online survey (Google Forms) used a 12-item questionnaire (4 demographics; 8 awareness domains) rated on a five-point Likert scale. Participants were recruited via snowball sampling. Descriptive statistics, independent-samples t-tests, and one-way analysis of variance with Tukey post hoc tests were applied (p < 0.05).

RESULTS: A total of 257 dentists participated. Mean domain scores ranged from 2.97 to 3.13; awareness of ethical issues was highest (3.13 ± 1.44) and perception of encryption lowest (2.97 ± 1.45). No significant gender differences were observed. University hospital dentists reported significantly higher awareness of ethical issues and a greater perceived need for anonymization than other institution types (p < 0.05). For awareness of Personal Data Protection Law (KVKK), scores were higher in university hospitals and private dental polyclinics than in private practices and public hospitals (p < 0.05). Professional experience was associated with differences in perception of encryption, awareness of personal data protection law, awareness of ethical issues, and perceived AI ethical risks (p < 0.05); perception of data security and awareness of big data security did not differ significantly (p > 0.05).

CONCLUSION: Dentists demonstrated varying levels of awareness across domains, with higher awareness reported in academic settings. Experience-related patterns differed by domain, indicating the need for focused educational strategies addressing legal and ethical aspects of AI-supported dental practice.

PMID:41637082 | DOI:10.2340/aos.v85.45419

JAMA Netw Open. 2026 Feb 2;9(2):e2555855. doi: 10.1001/jamanetworkopen.2025.55855.

ABSTRACT

IMPORTANCE: Failure to rescue (FTR), defined as postoperative mortality among patients with treatable complications, is a recognized patient safety concern. FTR reflects institutional capacity for timely management of deterioration and has been proposed as a quality indicator less dependent on baseline complication risk. Evidence on systematic hospital-level variation outside the US remains limited.

OBJECTIVE: To estimate national postoperative FTR rates, quantify between-hospital variation, and identify hospitals with better- or worse-than-expected performance using risk-standardized mortality ratios (RSMRs).

DESIGN, SETTING, AND PARTICIPANTS: This retrospective cohort study conducted in Switzerland applied the Agency for Healthcare Research and Quality (AHRQ) patient safety indicator 04 (PSI04) definition to administrative hospital data to all acute-care hospitals in Switzerland from January 2019 to December 2023. Participants included surgical inpatients with at least 1 PSI04-defined complication (ie, deep vein thrombosis and/or pulmonary embolism, pneumonia, sepsis, shock and/or cardiac arrest, and gastrointestinal hemorrhage and/or ulcer). Hospital-level variation was assessed using multilevel logistic regression with hospital random intercepts and summarized with RSMRs. Alternative models were estimated to explore the stability of results.

EXPOSURE: Acute care hospitalization.

MAIN OUTCOMES AND MEASURES: In-hospital mortality following eligible complications, expressed as crude FTR rates and RSMRs. The intraclass correlation coefficient quantified systematic performance variation.

RESULTS: Among 41 506 inpatients undergoing surgery with PSI04-defined complications (mean [SD] age, 67.6 [14.8] years; 24 692 [59.5%] men), 7310 in-hospital deaths occurred. The crude national FTR rate was 18.07 (95% CI, 17.66-18.50) of 100 admissions. In 61 hospitals with at least 100 cases, adjusted odds ratio for death varied between the lowest- and highest-performing hospitals from 0.56 (95% CI, 0.38-0.80) to 1.75 (95% CI, 1.59-1.92). Hospital-level variance was 0.114 (intraclass correlation coefficient, 0.034; 95% CI, 0.020-0.055). An estimated 1045 of 7114 observed FTR deaths (14.7%) within the hospital sample were attributable to below-average hospital performance. Five hospitals (8.2%) performed significantly better than expected, 42 (68.9%) as expected, and 14 (23.0%) substantially worse than expected based on RSMR 95% CIs. Poorer performance clustered in medium- and high-volume hospitals. Alternative regression models confirmed stability of results.

CONCLUSIONS AND RELEVANCE: In this cross-sectional study of FTR, nearly 1 in 5 patients undergoing surgery who experienced serious complications died, with substantial between-hospital variation. Multilevel modeling indicated that institutional performance accounted for 1045 potentially avoidable deaths. These findings support FTR as an international patient safety indicator and highlight the need to investigate organizational determinants of variation to inform system-level improvement strategies.

PMID:41637075 | DOI:10.1001/jamanetworkopen.2025.55855

JAMA Netw Open. 2026 Feb 2;9(2):e2555875. doi: 10.1001/jamanetworkopen.2025.55875.

ABSTRACT

IMPORTANCE: Before elective surgery, direct oral anticoagulants (DOACs) are discontinued following a standardized protocol. However, this could result in insufficient lowering of DOAC levels that could increase bleeding risk.

OBJECTIVE: To estimate the proportion of patients with elevated DOAC levels at the time of elective surgery, evaluate factors associated with DOAC levels, and examine associated blood loss.

DESIGN, SETTING, AND PARTICIPANTS: This cohort study (DOAC Level Prior to Incision [DALI]) assessed adult patients prescribed a DOAC (apixaban, dabigatran, or rivaroxaban) for any indication and at any dose, undergoing an elective procedure requiring DOAC interruption between May 27, 2018, and February 25, 2024, at 2 Dutch hospitals.

EXPOSURE: Standardized interruption protocol (1 day before moderate- and 2 days before high bleeding-risk procedures) with interruption adjustments for the patient’s kidney function.

MAIN OUTCOMES AND MEASURES: Blood was drawn immediately before surgery to determine DOAC levels (by liquid chromatography-mass spectrometry). Proportions of preoperative DOAC levels of 30 ng/mL or higher and their 95% CIs were estimated, stratified by DOAC type and surgical bleeding risk. Factors associated with DOAC levels were identified through multivariable linear regression. Surgical blood loss and 30-day postoperative complications were described according to DOAC concentrations.

RESULTS: The study was terminated after including 257 patients (100 receiving apixaban, 100 receiving rivaroxaban, and 57 receiving dabigatran due to the slow inclusion rate of those receiving dabigatran; median [IQR] age, 72 [66-78] years; 173 male [67%]); 212 patients (82%) underwent a high bleeding-risk operation. Preprocedural DOAC levels were 30 ng/mL or higher in 7.6% (95% CI, 4.9%-11.6%) of patients. Dabigatran and rivaroxaban had similar proportions, whereas 13.1% (95% CI, 7.8%-21.2%) of patients treated with apixaban had levels of 30 ng/mL or higher. Treatment with apixaban, decreased kidney function, and a shorter interruption time were associated with higher levels. Surgical blood loss (median [range], 0 [0-4250] mL) was not associated with DOAC levels. Twelve patients (4.7%; 95% CI, 2.7%-8.0%), who all had DOAC levels less than 30 ng/mL, experienced major bleeding.

CONCLUSIONS AND RELEVANCE: In this cohort study, most patients following the current protocol had DOAC levels less than 30 ng/mL, although the proportion of patients with elevated levels was higher for apixaban. Preoperative DOAC levels were not associated with blood loss during surgery.

PMID:41637074 | DOI:10.1001/jamanetworkopen.2025.55875

JAMA Netw Open. 2026 Feb 2;9(2):e2555896. doi: 10.1001/jamanetworkopen.2025.55896.

ABSTRACT

IMPORTANCE: Equity, diversity, and inclusion (EDI) initiatives are politically polarizing and increasingly adopted in the health care setting. Their broader impact across different health care career types, career stages, and various levels of education remains largely unknown.

OBJECTIVE: To assess EDI programs and their associated outcomes within health care institutions.

DATA SOURCES: A Preferred Reporting Items for Systematic Reviews and Meta-analyses (PRISMA) 2020-compliant systematic review searching PubMed, Scopus, Web of Science, CINAHL, and PsychINFO databases from January 2010 to December 2023.

STUDY SELECTION: Two independent reviewers screened studies that assessed EDI programs or policies in health care institutions.

DATA EXTRACTION AND SYNTHESIS: Programs were categorized based on reported outcomes, including participant satisfaction, increased awareness of EDI-related topics, increases in the proportion of underrepresented minority individuals within medical education or the health care workforce, and overall program impact. Odds ratios (ORs) were pooled using a random-effects model. Analyses followed Preferred Reporting Items for Systematic Reviews and Meta-Analyses reporting guidelines. Analysis was conducted June 2025.

MAIN OUTCOMES AND MEASURES: Outcome measures included the proportion of diversity among the workforce, employee and patient satisfaction, and the proportion of employees recruited and retained after program implementation.

RESULTS: In total, 43 studies incorporating more than 15 000 individuals involved in EDI programs were included. Interventions were multifaceted, including 14 career advancement and training programs, 16 diversity representation programs, 11 academia and research support initiatives, and the growth of 2 pipeline programs. Furthermore, interventions demonstrated consistent improvement in EDI initiatives, with perceived benefit in promoting underrepresented minority populations. Findings from the meta-analysis of 2 studies showed that minority representation in competitive medical residencies increased after implementation of 2 EDI interventions (OR, 1.73; 95% CI, 1.21-2.47). Among the 43 studies included in the Joanna Briggs Institute assessment of methodological quality, 7 (16.3%) were rated as high quality, 20 (46.5%) as moderate quality, and 16 (37.2%) as low quality.

CONCLUSIONS AND RELEVANCE: In this systematic review and meta-analysis of EDI initiatives in health care institutions, programs were associated with an increased workforce diversity. These findings support the continued use of EDI initiatives to promote a more inclusive and equitable health care culture.

PMID:41637073 | DOI:10.1001/jamanetworkopen.2025.55896