Category: Statistics

Cardiovasc Diabetol. 2025 Feb 18;24(1):80. doi: 10.1186/s12933-025-02640-9.

ABSTRACT

BACKGROUND: Digitalization and big health system data open new avenues for targeted prevention and treatment strategies. We aimed to develop and validate prediction models for stroke and myocardial infarction (MI) in patients with type 2 diabetes based on routinely collected high-dimensional health insurance claims and compared predictive performance of traditional regression with state-of-the-art machine learning including deep learning methods.

METHODS: We used German health insurance claims from 2014 to 2019 with 287 potentially relevant literature-derived variables to predict 3-year risk of MI and stroke. Following a train-test split approach, we compared the performance of logistic methods with and without forward selection, LASSO-regularization, random forests (RF), gradient boosting (GB), multi-layer-perceptrons (MLP) and feature-tokenizer transformers (FTT). We assessed discrimination (Areas Under the Precision-Recall and Receiver-Operator Curves, AUPRC and AUROC) and calibration.

RESULTS: Among n = 371,006 patients with type 2 diabetes (mean age: 67.2 years), 3.5% (n = 13,030) had MIs and 3.4% (n = 12,701) strokes. AUPRCs were 0.035 (MI) and 0.034 (stroke) for a null model, between 0.082 (MLP) and 0.092 (GB) for MI, and between 0.061 (MLP) and 0.073 (GB) for stoke. AUROCs were 0.5 for null models, between 0.70 (RF, MLP, FTT) and 0.71 (all other models) for MI, and between 0.66 (MLP) and 0.69 (GB) for stroke. All models were well calibrated.

CONCLUSIONS: Discrimination performance of claims-based models reached a ceiling at around 0.09 AUPRC and 0.7 AUROC. While for AUROC this performance was comparable to existing epidemiological models incorporating clinical information, comparison of other, potentially more relevant metrics, such as AUPRC, sensitivity and Positive Predictive Value was hampered by lack of reporting in the literature. The fact that machine learning including deep learning methods did not outperform more traditional approaches may suggest that feature richness and complexity were exploited before the choice of algorithm could become critical to maximize performance. Future research might focus on the impact of different feature derivation approaches on performance ceilings. In the absence of other more powerful screening alternatives, applying transparent regression-based models in routine claims, though certainly imperfect, remains a promising scalable low-cost approach for population-based cardiovascular risk prediction and stratification.

PMID:39966813 | DOI:10.1186/s12933-025-02640-9

BMC Musculoskelet Disord. 2025 Feb 18;26(1):163. doi: 10.1186/s12891-025-08413-5.

ABSTRACT

OBJECTIVE: This study aimed to determine the optimal index for distinguishing ankylosing spondylitis (AS) from non-AS by employing texture analysis of bone marrow edema (BME) in lumbar spine MR images.

METHODS: We conducted a retrospective analysis, involving patients meeting specific criteria with positive BME signs in lumbar spine MRI. We compared 72 cases (78 lesions) from the AS group with 67 cases (84 lesions) from the non-AS group. Image acquisition was single-blind, and we defined the region of interest (ROI) at the lumbar spine’s maximal BME level using ImageJ software. Texture analysis parameters were extracted from Gray Level Histogram(GLH) and Gray-Level Co-occurrence Matrix(GLCM) of STIR and T2WI sequences in both groups. We generated Receiver Operating Characteristic(ROC) curves based on statistically significant parameters and calculated the area under the curve (AUC).

RESULTS: In BME STIR GLH analysis, AS group had higher Mean, Mode, Min, and Skew parameters than the non-AS group (p < 0.001), with Min exhibiting the highest diagnostic efficacy (AUC = 0.768). T2WI GLH analysis showed that only Min was significantly higher in the AS group (p = 0.014,AUC = 0.612). Analysis of BME zones in the STIR GLCM revealed significant differences in ASM and Ent parameters between the AS and non-AS groups, with ASM displaying the highest diagnostic accuracy (p < 0.001,AUC = 0.656). For T2WI GLCM analysis, all four parameters (ASM, Cor, IDM, and Ent) were significantly different in the AS group (p < 0.001), with ASM demonstrating the highest diagnostic accuracy (AUC = 0.731).

CONCLUSIONS: Lumbar BME texture analysis effectively distinguishes AS from non-AS, with significant variations in multiple parameter values. The STIR GLH parameter Min provides the highest diagnostic accuracy.

PMID:39966811 | DOI:10.1186/s12891-025-08413-5

BMC Public Health. 2025 Feb 18;25(1):667. doi: 10.1186/s12889-025-21970-7.

ABSTRACT

BACKGROUND: Community-acquired pneumonia (CAP) significantly affects elderly patients, leading to high morbidity and mortality rates. This study investigates the interplay between mental health, nutritional status, and comorbidities in determining the prognosis of elderly patients with CAP.

METHODS: A retrospective cohort study was conducted with 455 patients aged 75 and older who were hospitalized for CAP. Clinical data, including demographic information, comorbidities, and laboratory results, were collected. The WHO-5 Well-Being Index (WHO-5), Mini Nutritional Assessment Short Form (MNA-SF), and Charlson Comorbidity Index (CCI) were utilized to assess mental health, nutritional status, and comorbidity burden. Statistical analyses included logistic regression, Kaplan-Meier survival analysis, and mediation analyses.

RESULTS: The study found that the 28-day mortality rate was 9.67%, while the 90-day mortality rate reached 12.31%. Spearman’s correlation analysis revealed significant positive correlations between the WHO-5 Well-Being Index and MNA-SF scores (r = 0.560) and albumin levels (r = 0.245), while negative correlations were observed with CCI (r = -0.202) and C-reactive protein levels (r = -0.242). Logistic regression analysis indicated that comorbidity, malnutrition, lower well-being, CAP severity, and mechanical ventilation are significant predictors of 28-day and 90-day mortality. Kaplan-Meier survival analysis demonstrated statistically significant differences in cumulative survival among various well-being groups. Multiple mediation analyses showed that mental well-being and nutritional status significantly mediated the association between CCI and 28-day and 90-day mortality.

CONCLUSION: This study emphasizes the critical roles of mental health, nutritional status, and comorbidities in the prognosis of elderly patients with CAP. Integrating these factors into clinical assessments may provide insights to inform management strategies, potentially improving patient outcomes and reducing mortality rates in this vulnerable population.

PMID:39966810 | DOI:10.1186/s12889-025-21970-7

J Ethnobiol Ethnomed. 2025 Feb 18;21(1):8. doi: 10.1186/s13002-025-00753-9.

ABSTRACT

BACKGROUND: Throughout history, plant resources have played a crucial role in human society. After addressing fundamental needs such as food and shelter, humans have sought out plants for medicinal purposes to alleviate various health issues. The utilization of plant resources for diverse applications, including traditional herbal medicine, is integral to the rich cultural heritage and lifestyle of the communities in southwest Ethiopia. However, despite the existence of numerous indigenous traditional medicinal plants, the ethnobotanical knowledge surrounding these resources in the Bita district remains largely unexplored. Consequently, this study aimed to document and analyze the traditional medicinal plants, along with the associated customs and knowledge utilized by the local population.

METHODS: Between June 2024 and Pagume (the 13th month unique to Ethiopia) of the same year, a combination of semistructured interviews, in-person meetings, group discussions, and guided field trips was employed to collect quantitative ethnobotanical data. A total of 136 informants, comprising 104 men and 32 women, participated in the interviews to provide insights into ethnobotanical practices. The research utilized several quantitative methodologies, including the informant consensus factor (ICF), fidelity level (FL), plant part value, preference ranking, and direct matrix ranking. Additionally, various statistical analyses were conducted, including independent t tests, one-way ANOVA, correlation, and regression, utilizing R to assess and compare the ethnobotanical knowledge across different groups of informants.

RESULT: A total of 122 species of traditional medicinal plants, belonging to 104 genera and 53 different plant families, were documented in this study. The Asteraceae family was the most frequently cited, comprising 12 species, making it the largest family identified. This was followed by Lamiaceae with eight species, Solanaceae with eight species, Rubiaceae with seven species, Euphorbiaceae with six species, Cucurbitaceae with five species, and Fabiaceae with four species. The plant parts most commonly utilized in traditional remedies were leaves and roots, with the predominant method of preparation being crushing. Notably, the average number of medicinal plants reported by participants varied significantly across different demographics, including gender, age groups, educational levels, and experience (P < 0.05).

CONCLUSION: The study area boasts a diverse range of potential medicinal plants and the associated indigenous knowledge. To mitigate the increasing anthropogenic threats and ensure the preservation of these plants and their related knowledge, it is crucial to implement effective conservation strategies and responsible usage. Furthermore, the medicinal properties of these plants should be validated through scientific experimentation to effectively combine local knowledge with modern medicine.

PMID:39966803 | DOI:10.1186/s13002-025-00753-9

BMC Musculoskelet Disord. 2025 Feb 18;26(1):164. doi: 10.1186/s12891-025-08414-4.

ABSTRACT

BACKGROUND: Ankle fractures are the third most common fractures, often requiring surgical intervention to restore function and mobility. Understanding trends in ankle fracture surgeries is essential for optimizing treatment strategies and improving patient outcomes. The aim of this study is to provide a comprehensive analysis of ankle fracture surgeries in Sweden in order to highlight changes in demography and trends in surgical procedures.

METHODS: Utilizing data from the National Patient Register, we conducted an observational population-based study of ankle fracture surgeries performed in Sweden between 2008 and 2022. Patients aged 15 years and above who underwent ankle fracture surgery were included in the analysis. Surgical procedures were identified using NOMESCO codes specific to ankle fractures. Demographic trends, surgical procedures and incidence rates were analyzed using descriptive statistics, incidence calculations and regression analyses.

RESULTS: Women accounted for 55% of surgeries (p = 0.022), with a significant proportion (47%) occurring in individuals aged 65 and above. Although the overall incidence of ankle surgeries decreased by 6% (p = 0.008), notable age-specific trends emerged, including a decrease in surgeries among younger adults and an increase among the elderly. Plate and screw fixation remained the most commonly employed surgical technique, with a 21% increase in usage (p < 0.001), while the use of external fixation and intramedullary nailing increased significantly by 123% and 69%, respectively (both p < 0.001). Conversely, the use of cerclage and/or pin fixation decreased by 74% (p < 0.001) over the study period.

CONCLUSIONS: Our study of open source data shows current trends in surgically treated ankle fractures in Sweden, highlighting a decreased incidence overall, notable shifts between age groups and several trends in surgical procedures. Despite limitations inherent to retrospective observational studies, such as the inability to establish causal relationships, our findings contribute to the understanding of ankle fracture management trends, highlighting areas for further investigation and improvement in orthopedic care.

PMID:39966800 | DOI:10.1186/s12891-025-08414-4

BMC Cancer. 2025 Feb 18;25(1):292. doi: 10.1186/s12885-025-13713-z.

NO ABSTRACT

PMID:39966798 | DOI:10.1186/s12885-025-13713-z

BMC Oral Health. 2025 Feb 18;25(1):251. doi: 10.1186/s12903-025-05622-1.

ABSTRACT

BACKGROUND: Free gingival graft (FGG) is considered as a well-established periodontal surgical technique to achieve sufficient keratinized tissue width and thickness and subsequently enhance gingival health and stability. This bibliometric research aims to reveal research focuses and trends about FGG.

METHODS: Articles published on FGG were retrospectively retrieved from the Web of Science Core Collection database from 2000 to 2023. Statistical and visual analyses were performed to characterize their quantity, journals, countries and regions, institutions, authors and keywords by CiteSpace software.

RESULTS: The top 100 most cited articles comprised 80 original research papers and 20 reviews, with an average citation count of 56. Notably, 75% of these works (n = 75) were classified in the Q1 category of the Journal Citation Reports (JCR). The most influential article authored by Dr. Cairo F. in 2014 has received 251 citations. Specifically, 27 high-level papers published in Journal of Periodontology accounted for 1,849 citations. The United States with 30 articles published and the University of Michigan with 11 articles were the most productive country and institution, respectively. Prof. Wang Hom-lay published 6 articles with a total of 608 citations. Additionally, collagen matrix and dental implants have garnered significant attention over the past decades.

CONCLUSION: Our analysis offers a comprehensive overview and in-depth analysis of the future development trends and potential research directions of FGG, which can inspire both clinical and scientific researchers.

PMID:39966796 | DOI:10.1186/s12903-025-05622-1

BMC Med Inform Decis Mak. 2025 Feb 18;25(1):89. doi: 10.1186/s12911-024-02731-9.

ABSTRACT

BACKGROUND: The exponential growth in patient data collection by healthcare providers, governments, and private industries is yielding large and diverse datasets that offer new insights into critical medical questions. Leveraging extensive computational resources, Machine Learning and Artificial Intelligence are increasingly utilized to address health-related issues, such as predicting outcomes from Electronic Health Records and detecting patterns in multi-omics data. Despite the proliferation of medical devices based on Artificial Intelligence, data accessibility for research is limited due to privacy concerns. Efforts to de-identify data have met challenges in maintaining effectiveness, particularly with large datasets. As an alternative, synthetic data, that replicate main statistical properties of real patient data, are proposed. However, the lack of standardized evaluation metrics complicates the selection of appropriate synthetic data generation methods. Effective evaluation of synthetic data must consider resemblance, utility and privacy, tailored to specific applications. Despite available metrics, benchmarking efforts remain limited, necessitating further research in this area.

RESULTS: We present SynthRO (Synthetic data Rank and Order), a user-friendly tool for benchmarking health synthetic tabular data across various contexts. SynthRO offers accessible quality evaluation metrics and automated benchmarking, helping users determine the most suitable synthetic data models for specific use cases by prioritizing metrics and providing consistent quantitative scores. Our dashboard is divided into three main sections: (1) Loading Data section, where users can locally upload real and synthetic datasets; (2) Evaluation section, in which several quality assessments are performed by computing different metrics and measures; (3) Benchmarking section, where users can globally compare synthetic datasets based on quality evaluation.

CONCLUSIONS: Synthetic data mitigate concerns about privacy and data accessibility, yet lacks standardized evaluation metrics. SynthRO provides an accessible dashboard helping users select suitable synthetic data models, and it also supports various use cases in healthcare, enhancing prognostic scores and enabling federated learning. SynthRO’s accessible GUI and modular structure facilitate effective data evaluation, promoting reliability and fairness. Future developments will include temporal data evaluation, further broadening its applicability.

PMID:39966793 | DOI:10.1186/s12911-024-02731-9

BMC Oral Health. 2025 Feb 18;25(1):254. doi: 10.1186/s12903-025-05624-z.

ABSTRACT

BACKGROUND: Analyzing the morphological growth changes upon mandibular advancement between Twin Block (TB) and Functional Regulator II (FR2) in Class II patients involves measuring the condylar and mandibular changes in terms of linear and volumetric measurements, as well as the distances registered after superimposition. Correspondingly, 3D digital assessment with the benefit of artificial intelligence tools was applied, and color-coded distance maps were descriptively interpreted to visualize the direction of growth.

SUBJECTS AND METHODS: Twenty patients aged 9 to 12 years with skeletal Class II malocclusion with overjet (5-9 mm) were enrolled into two groups: Group 1 (TB) and Group 2 (FR2). CBCT radiographs were taken at T0 (prior to treatment) and T1 (at the end of effective treatment; 7-11 months). 3D UX-Net algorithm was used to extract the 3D models and obtain the volumes. Visual Basic software was employed to develop a method for calculating the 3D distances and angles. 3D superimposition methods through Viewbox software were utilized to analyze growth direction following the treatments, using color-coded distance map visualizations.

RESULTS: The average volume increase of the mandible, right and left condyle was 2666.00 ± 1469.16 mm³, 207.70 ± 125.28 mm³, and 216.90 ± 121.65 mm³, respectively, after TB treatment, and 3316.10 ± 1552.29 mm³, 208.90 ± 116.85 mm³, and 193.65 ± 114.25 mm³, respectively, after FR2 treatment. The mean absolute distances (MAD) for the mandible, right, and left condyle from T1 to T0 were in TB group 1.238 ± 0.757 mm, 2.248 ± 0.346 mm, 2.267 ± 0.593 mm, respectively, and in FR2 group 1.417 ± 0.550 mm, 2.616 ± 1.063 mm, 2.475 ± 0.916 mm, respectively. The visualization of color-coded maps revealed a superior-posterior growth direction of the condyle measuring within 3 mm after the treatment.

CONCLUSIONS: Although mandibular and condylar volumes increased in each group when comparing T0 to T1, indicating the efficacy of both appliances in enhancing growth, no clinically or statistically significant differences were observed between the groups in terms of volumetric, linear, and 3D superimposition changes. However, color-coded distance maps demonstrated a consistent pattern of posterior-superior relocation of the condyles in most cases after treatment when comparing T1 to T0, whereas the back of the ramus in more than half of the cases has moved backwards. With no statistical differences between the studied groups.

PMID:39966790 | DOI:10.1186/s12903-025-05624-z

BMC Health Serv Res. 2025 Feb 18;25(1):274. doi: 10.1186/s12913-025-12426-0.

ABSTRACT

BACKGROUND: Nurses have a crucial role in recognizing potential dangers to patient safety during direct patient care. However, there are occasions when they opt to stay quiet instead of voicing their concerns, ultimately endangering patients. The underlying reasons for this silence pose a significant challenge in the nursing industry. While it’s known that workplace conditions contribute to this silence, the specific ways in which this occurs need to be clarified. This research aimed to investigate the factors that lead nurses to remain silent about patient safety, specifically focusing on environmental influences in Ardabil City.

METHODS: This descriptive and analytical study examined 630 nurses, including head nurses, from five medical training centers in Ardabil, northwest Iran. The primary instruments used in the research included the Employee Silence about Patient Safety Questionnaire, the Practice Environment Scale of the Nursing Work Index (PES-NWI), measures of psychological safety, leader-member exchange (LMX) assessments, and evaluations of professional discrimination. Data analysis was performed using SPSS version 16.0, applying descriptive statistics, t-tests, ANOVA, Pearson’s correlation coefficient, and hierarchical regression analysis.

RESULTS: The average score for employee silence regarding patient safety was 2.62, with a standard deviation of 0.98. Several factors were found to be significant predictors of employees’ silence about patient safety, including nurse participation in hospital affairs (ß = 0.196, p = 0.002), nurse’s role in the quality of care (ß = -0.352, p < 0.001), staff and facility adequacy (ß = 0.156, p = 0.001), communication with the physician (ß = -0.105, p = 0.015), LMX (ß = -0.284, p < 0.001), and job position (ß = -0.093, p = 0.018).

CONCLUSION: This study has identified several critical factors influencing employee silence regarding patient safety, including nurse involvement in hospital decision-making, the quality of care, communication with physicians, and job positions. To address these challenges, healthcare organizations must implement standardized protocols, improved communication channels, regular safety training, and the integration of advanced technologies. Increasing nurse participation in decision-making processes and strengthening communication between nursing staff and physicians can foster a culture of openness that encourages employees to voice their safety concerns. By creating an environment where staff feel supported and empowered to speak up, healthcare facilities can enhance patient safety and improve overall care quality, ultimately leading to a safer working environment for healthcare professionals.

PMID:39966779 | DOI:10.1186/s12913-025-12426-0