Category: Statistics

J Am Stat Assoc. 2023;118(544):2315-2328. doi: 10.1080/01621459.2022.2044824. Epub 2022 Mar 27.

ABSTRACT

In this paper, we leverage over-parameterization to design regularization-free algorithms for the high-dimensional single index model and provide theoretical guarantees for the induced implicit regularization phenomenon. Specifically, we study both vector and matrix single index models where the link function is nonlinear and unknown, the signal parameter is either a sparse vector or a low-rank symmetric matrix, and the response variable can be heavy-tailed. To gain a better understanding of the role played by implicit regularization without excess technicality, we assume that the distribution of the covariates is known a priori. For both the vector and matrix settings, we construct an over-parameterized least-squares loss function by employing the score function transform and a robust truncation step designed specifically for heavy-tailed data. We propose to estimate the true parameter by applying regularization-free gradient descent to the loss function. When the initialization is close to the origin and the stepsize is sufficiently small, we prove that the obtained solution achieves minimax optimal statistical rates of convergence in both the vector and matrix cases. In addition, our experimental results support our theoretical findings and also demonstrate that our methods empirically outperform classical methods with explicit regularization in terms of both ℓ₂-statistical rate and variable selection consistency.

PMID:38550788 | PMC:PMC10977662 | DOI:10.1080/01621459.2022.2044824

Front Digit Health. 2024 Mar 14;6:1365345. doi: 10.3389/fdgth.2024.1365345. eCollection 2024.

ABSTRACT

OBJECTIVE: This umbrella review aims to ascertain the extent to which immersive Virtual Reality (VR) and Augmented Reality (AR) technologies improve specific competencies in healthcare professionals within medical education and training, in contrast to traditional educational methods or no intervention.

METHODS: Adhering to PRISMA guidelines and the PICOS approach, a systematic literature search was conducted across major databases to identify studies examining the use of VR and AR in medical education. Eligible studies were screened and categorized based on the PICOS criteria. Descriptive statistics and chi-square tests were employed to analyze the data, supplemented by the Fisher test for small sample sizes or specific conditions.

ANALYSIS: The analysis involved cross-tabulating the stages of work (Development and Testing, Results, Evaluated) and variables of interest (Performance, Engagement, Performance and Engagement, Effectiveness, no evaluated) against the types of technologies used. Chi-square tests assessed the associations between these categorical variables.

RESULTS: A total of 28 studies were included, with the majority reporting increased or positive effects from the use of immersive technologies. VR was the most frequently studied technology, particularly in the “Performance” and “Results” stages. The chi-square analysis, with a Pearson value close to significance (p = 0.052), suggested a non-significant trend toward the association of VR with improved outcomes.

CONCLUSIONS: The results indicate that VR is a prevalent tool in the research landscape of medical education technologies, with a positive trend toward enhancing educational outcomes. However, the statistical analysis did not reveal a significant association, suggesting the need for further research with larger sample sizes. This review underscores the potential of immersive technologies to enhance medical training yet calls for more rigorous studies to establish definitive evidence of their efficacy.

PMID:38550715 | PMC:PMC10973128 | DOI:10.3389/fdgth.2024.1365345

Pain Res Manag. 2024 Mar 20;2024:9989997. doi: 10.1155/2024/9989997. eCollection 2024.

ABSTRACT

BACKGROUND: Patients undergoing breast surgery are at risk of severe postoperative pain. Several opioid-sparing strategies exist to alleviate this condition. Regional anesthesia has long been a part of perioperative pain management for these patients.

AIM: This randomized study examined the benefits of interpectoral and pectoserratus plane block (IPP/PSP), also known as pectoralis nerve plain block, compared with advanced local anesthetic infiltration.

METHODS: We analyzed 57 patients undergoing partial mastectomy with sentinel node dissection. They received either an ultrasound-guided IPP/PSP block performed preoperatively by an anesthetist or local anesthetic infiltration performed by the surgeon before and during the surgery.

RESULTS: Pain measured with the numerical rating scale (NRS) indicated no statistically significant difference between the groups (IPP/PSP 1.67 vs. infiltration 1.97; p value 0.578). Intraoperative use of fentanyl was significantly lower in the IPP/PSP group (0.18 mg vs 0.21 mg; p value 0.041). There was no statistically significant difference in the length of stay in the PACU (166 min vs 175 min; p value 0.51). There were no differences in reported postoperative nausea and vomiting (PONV) between the groups. The difference in postoperative use of oxycodone in the PACU (p value 0.7) and the use of oxycodone within 24 hours postoperatively (p value 0.87) was not statistically significant.

CONCLUSIONS: Our study showed decreased intraoperative opioid use in the IPP/PSP group and no difference in postoperative pain scores up to 24 hours. Both groups reported low postoperative pain scores. This trial is registered with NCT04824599.

PMID:38550709 | PMC:PMC10977337 | DOI:10.1155/2024/9989997

Front Psychol. 2024 Mar 14;15:1355752. doi: 10.3389/fpsyg.2024.1355752. eCollection 2024.

ABSTRACT

INTRODUCTION: During the pandemic, young people experienced a general increase in stress levels in their home and school environments and in their relationships with peers and family, largely due to restrictions on freedom of movement and social isolation. The ability to identify sources of stress and respond positively to them, using both personal and environmental resources, seems to be key to maintaining an acceptable level of well-being. This study investigates the association between alexithymic traits, self-perceived well-being, and self-representations in adolescents as expressed via narrative metaphors during the COVID-19 epidemic.

METHODS: The sample comprised 229 Italian adolescents (51.1% females, mean age = 16.64). The research design was based on an exploratory, parallel, mixed-method approach. A semi-structured online interview was used as the major data gathering tool including both standardized quantitative questionnaire and open-ended questions. Data were analyzed by means of descriptive statistics, quantitative textual analysis and multidimensional co-word correspondence analysis.

RESULTS: Main findings reveal a general low level of perceived well-being associated with alexithymia, affecting adolescents’ lexical choices for their metaphors. Alexithymia-related low levels of well-being correspond to metaphors in which confusion and overpowering emotions predominate. Vivid pictures indicating vitality and a bright view on the future is often correlated with high levels of well-being.

DISCUSSION: Overall, these novel findings appear to show an interactive effect of perceived well-being and alexithymia on adolescents’ ability to identify and describe their own condition. Furthermore, metaphors emerge as powerful tools for investigating well-being in adolescents since closely related to inner states.

PMID:38550637 | PMC:PMC10973111 | DOI:10.3389/fpsyg.2024.1355752

Vet Med (Praha). 2024 Feb 27;69(2):42-51. doi: 10.17221/49/2023-VETMED. eCollection 2024 Feb.

ABSTRACT

Heart rate variability analyses using Poincaré plots can be useful for evaluating the autonomic nervous system function. However, the interpretation of the quantitative indicators of Poincaré plots remains controversial. Thus, few studies have verified the effectiveness of the quantitative indicators in veterinary medicine. This study aimed to verify the reliability of Poincaré plot indicators using pharmacological models in dogs. Four healthy beagles were used in this study. Each dog was treated with propranolol, atropine, and propranolol-atropine to block the sympathetic, parasympathetic, and sympathetic-parasympathetic functions, respectively. The quantitative indicators of the Poincaré plots were calculated based on data from 300 electrocardiogram beats collected before and after the administration of each drug and statistically analysed. The quantitative indicators of the Poincaré plots, such as the standard deviation perpendicular to the major axis (SD1), standard deviation along the major axis (SD2), and SD1 × SD2, significantly decreased after the drug administration in both the parasympathetic and sympathetic-parasympathetic blockade models. However, no significant differences were observed in SD1/SD2 between the groups. The Poincaré plots reflected the changes in the autonomic nervous system of dogs. In dogs, SD1, SD2, and SD1 × SD2 can detect a state in which parasympathetic nerve activity is suppressed.

PMID:38550620 | PMC:PMC10966429 | DOI:10.17221/49/2023-VETMED

Front Cell Infect Microbiol. 2024 Mar 14;14:1340610. doi: 10.3389/fcimb.2024.1340610. eCollection 2024.

ABSTRACT

BACKGROUND: Combination antiretroviral therapy (ART) has transformed human immunodeficiency virus (HIV) infection in people with HIV (PWH). However, a chronic state of immune activation and inflammation is maintained despite achieving HIV suppression and satisfactory immunological recovery. We aimed to determine whether the plasma metabolomic profile of PWH on long-term suppressive ART and immunologically recovered approximates the normality by comparison with healthy controls with similar age and gender.

METHODS: We carried out a cross-sectional study in 17 PWH on long-term ART (HIV-RNA <50 copies/mL, CD4⁺ ≥500 cells/mm³, and CD4⁺/CD8⁺ ≥1) and 19 healthy controls with similar age and gender. Metabolomics analysis was performed by gas chromatography-mass spectrometry (GC-MS) and liquid chromatography-mass spectrometry (LC-MS). The statistical association analysis was performed by principal component analysis (PCA), partial least squares discriminant analysis (PLS-DA), and Generalized Linear Models (GLM) with a gamma distribution (log-link). Significance levels (p-value) were corrected for multiple testing (q-value).

RESULTS: PCA and PLS-DA analyses found no relevant differences between groups. Adjusted GLM showed 14 significant features (q-value<0.20), of which only three could be identified: lysophosphatidylcholine (LysoPC) (22:6) (q-value=0.148), lysophosphatidylethanolamine (LysoPE) (22:6) (q-value=0.050) and hydroperoxy-octadecatrienoic acid (HpOTrE)/dihydroperoxy-octadecatrienoic acid (DiHOTrE)/epoxy-octadecadienoic acid (EpODE) (q-value=0.136). These significant identified metabolites were directly correlated to plasma inflammatory biomarkers in PWH and negatively correlated in healthy controls.

CONCLUSION: PWH on long-term ART have a metabolomic profile that is almost normal compared to healthy controls. Nevertheless, residual metabolic alterations linked to inflammatory biomarkers persist, which could favor the development of age-related comorbidities among this population.

PMID:38550617 | PMC:PMC10972849 | DOI:10.3389/fcimb.2024.1340610

Transact Mach Learn Res. 2023 Sep;2023:https://openreview.net/forum?id=oud7Ny0KQy.

ABSTRACT

The ubiquity of missing values in real-world datasets poses a challenge for statistical inference and can prevent similar datasets from being analyzed in the same study, precluding many existing datasets from being used for new analyses. While an extensive collection of packages and algorithms have been developed for data imputation, the overwhelming majority perform poorly if there are many missing values and low sample sizes, which are unfortunately common characteristics in empirical data. Such low-accuracy estimations adversely affect the performance of downstream statistical models. We develop a statistical inference framework for regression and classification in the presence of missing data without imputation. Our framework, RIFLE (Robust InFerence via Low-order moment Estimations), estimates low-order moments of the underlying data distribution with corresponding confidence intervals to learn a distributionally robust model. We specialize our framework to linear regression and normal discriminant analysis, and we provide convergence and performance guarantees. This framework can also be adapted to impute missing data. In numerical experiments, we compare RIFLE to several state-of-the-art approaches (including MICE, Amelia, MissForest, KNN-imputer, MIDA, and Mean Imputer) for imputation and inference in the presence of missing values. Our experiments demonstrate that RIFLE outperforms other benchmark algorithms when the percentage of missing values is high and/or when the number of data points is relatively small. RIFLE is publicly available at https://github.com/optimization-for-data-driven-science/RIFLE.

PMID:38550611 | PMC:PMC10977932

Front Immunol. 2024 Mar 14;15:1328297. doi: 10.3389/fimmu.2024.1328297. eCollection 2024.

ABSTRACT

BACKGROUND: Obesity and dyslipidemia, major global health concerns, have been linked to psoriasis, but previous studies faced methodological limitations and their shared genetic basis remains unclear. This study examines various obesity-related and lipidemic traits as potential contributors to psoriasis development, aiming to clarify their genetic associations and potential causal links.

METHODS: Summary statistics from genome-wide association studies (GWAS) conducted for obesity-related traits (body mass index (BMI), waist-to-hip ratio (WHR), and waist-to-hip ratio adjusted for the body mass index (WHR_adjBMI)) and lipidemic traits (high-density lipoprotein (HDL), LDL, triglyceride (TG), total Cholesterol (TC), apolipoprotein A1 (apoA1), apolipoprotein B (apoB), and apolipoprotein E (apoE)) and psoriasis, all in populations of European ancestry, were used. We quantified genetic correlations, identified shared loci and explored causal relationship across traits.

RESULTS: We found positive genetic correlation between BMI and psoriasis (r_g=0.22, p=2.44×10^-18), and between WHR and psoriasis (r_g=0.19, p=1.41×10^-12). We further found the positive genetic correlation between psoriasis and WHR_adjBMI(r_g=0.07, p=1.81×10^-2) the genetic correlation, in while the effect of BMI was controlled for. We identified 14 shared loci underlying psoriasis and obesity-related traits and 43 shared loci between psoriasis and lipidemic traits via cross-trait meta-analysis. Mendelian randomization (MR) supported the causal roles of BMI (IVW OR=1.483, 95%CI=1.333-1.649), WHR (IVW OR=1.393, 95%CI=1.207-1.608) and WHR_adjBMI (IVW OR=1.18, 95%CI=1.047-1.329) in psoriasis, but not observe any significant association between lipidemic traits and the risk of psoriasis. Genetic predisposition to psoriasis did not appear to affect the risk of obesity and lipidemic traits.

CONCLUSIONS: An intrinsic link between obesity-related traits and psoriasis has been demonstrated. The genetic correlation and causal role of obesity-related traits in psoriasis highlight the significance of weight management in both the prevention and treatment of this condition.

PMID:38550599 | PMC:PMC10972863 | DOI:10.3389/fimmu.2024.1328297

J Public Health Dent. 2024 Mar 28. doi: 10.1111/jphd.12615. Online ahead of print.

ABSTRACT

OBJECTIVES: Smoking is a major global health problem with serious systemic and oral consequences. This study aims at assessing the influence of smoking cigarettes and other types of smoked tobacco on oral conditions (OCs) using a representative sample of US adults.

METHODS: Pre-pandemic data from the National Health and Nutrition Examination Survey 2017-2020 were utilized, and 7840 adults aged ≥30 years were included in our analyses. Descriptive statistics, logistic, and negative binomial regression models were performed to assess the relationship between tobacco products and OCs including, tooth loss (TL), coronal (CC) and root caries (RC).

RESULTS: Overall, 16.29% of our sample were current cigarette smokers (CCS). TL (17.25%) and untreated RC (28.26%) were more evident among CCS. In the adjusted regression models, smoking cigarettes was associated with RC (AOR: 3.20, 95% CI; 2.02, 5.09), untreated CC (IRR: 3.08, 95% CI: 1.50, 6.31), and TL (IRR: 2.18, 95% CI: 1.64, 2.88). Regarding the type of used tobacco product in the past 5 days, cigarettes were the most common type (15.03%). The adjusted model indicated that e-cigarette smokers had the highest odds of untreated RC (AOR: 5.17, 95% CI: 2.19, 12.23) and the highest rate of TL (IRR: 2.21, 95% CI: 1.45, 3.35). Further, cigar smokers had the highest rate of teeth with untreated CC (IRR: 3.25, 95% CI: 1.46, 7.25).

CONCLUSIONS: Using tobacco products is associated with poor OCs. Dentists, being the primary oral health care providers, can play a crucial role in counseling and supporting smokers to quit as part of their routine dental examination.

PMID:38548675 | DOI:10.1111/jphd.12615

J Clin Pediatr Dent. 2024 Mar;48(2):173-180. doi: 10.22514/jocpd.2024.045. Epub 2024 Mar 3.

ABSTRACT

One of the most common congenital anomalies of the head and neck region is a cleft lip and palate. This retrospective case-control research aimed to compare the maxillary sinus volumes in individuals with bilateral cleft lip and palate (BCLP) to a non-cleft control group. The study comprised 72 participants, including 36 patients with BCLP and 36 gender and age-matched control subjects. All topographies were obtained utilizing Cone Beam Computed Tomography (CBCT) for diagnostic purposes, and 3D Dolphin software was utilized for sinus segmentation. Volumetric measurements were taken in cubic millimeters. No significant differences were found between the sex and age distributions of both groups. Additionally, there was no statistically significant difference observed between the BCLP group and the control group on the right and left sides (p > 0.05). However, the mean maxillary sinus volumes of BCLP patients (8014.26 ± 2841.03 mm³) were significantly lower than those of the healthy control group (11,085.21 ± 3146.12 mm³) (p < 0.05). The findings of this study suggest that clinicians should be aware of the lower maxillary sinus volumes in BCLP patients when planning surgical interventions. The utilization of CBCT and sinus segmentation allowed for precise measurement of maxillary sinus volumes, contributing to the existing literature on anatomical variations in BCLP patients.

PMID:38548647 | DOI:10.22514/jocpd.2024.045