Category: Statistics

IEEE Trans Cybern. 2022 Jun 10;PP. doi: 10.1109/TCYB.2022.3165879. Online ahead of print.

ABSTRACT

Fuzzy membership is an effective approach used in twin support vector machines (SVMs) to reduce the effect of noise and outliers in classification problems. Fuzzy twin SVMs (TWSVMs) assign membership weights to reduce the effect of outliers, however, it ignores the positioning of the input data samples and hence fails to distinguish between support vectors and noise. To overcome this issue, intuitionistic fuzzy TWSVM combined the concept of intuitionistic fuzzy number with TWSVMs to reduce the effect of outliers and distinguish support vectors from noise. Despite these benefits, TWSVMs and intuitionistic fuzzy TWSVMs still suffer from some drawbacks as: 1) the local neighborhood information is ignored among the data points and 2) they solve quadratic programming problems (QPPs), which is computationally inefficient. To overcome these issues, we propose a novel intuitionistic fuzzy weighted least squares TWSVMs for classification problems. The proposed approach uses local neighborhood information among the data points and also uses both membership and nonmembership weights to reduce the effect of noise and outliers. The proposed approach solves a system of linear equations instead of solving the QPPs which makes the model more efficient. We evaluated the proposed intuitionistic fuzzy weighted least squares TWSVMs on several benchmark datasets to show the efficiency of the proposed model. Statistical analysis is done to quantify the results statistically. As an application, we used the proposed model for the diagnosis of Schizophrenia disease.

PMID:35687632 | DOI:10.1109/TCYB.2022.3165879

PLoS One. 2022 Jun 10;17(6):e0269727. doi: 10.1371/journal.pone.0269727. eCollection 2022.

ABSTRACT

As a member state of the European Union, where vaccines against COVID-19 are available and affordable, Bulgaria reports the lowest immunization coverage and the most pronounced vaccine distrust. The present study aimed to assess the self-reported adverse reactions following COVID-19 vaccination as a possible tool to increase the trust in vaccines. A cross-sectional survey-based study, covering 761 vaccinated respondents, was conducted in Plovdiv (469 with an mRNA vaccine and 292 with an adenoviral vector vaccine). Descriptive statistics parametric and non-parametric methods were applied. Statistical significance was set at p<0.05. The median age of the respondents was 42 years, females (72.5%). At least one adverse reaction was reported in 89.9% of those immunized with mRNA vaccine and 93.8% in the adenoviral vector vaccine group (p>0.05). They were mild to moderate and resolved within several days. The levels of local reactions were comparable: 91.7% in those who received mRNA and 89.7% in those who received an adenoviral vector vaccine (p = 0.366). The most common types of systemic reactions were fatigue, headache, and muscle pains. An association was found between the systemic reactions and the type of vaccine administered: 59.7% in mRNA recipients and 89.4% in adenoviral vector vaccinees (p<0.001). None of the registered systemic reactions required medical attention. There were 3 reports of generalized urticaria after an mRNA and 2 after an adenoviral vector vaccine. The reported reactions are relatively high but expected and no adverse events have been reported that are not listed in the official Summary of Product Characteristics.

PMID:35687609 | DOI:10.1371/journal.pone.0269727

PLoS One. 2022 Jun 10;17(6):e0269802. doi: 10.1371/journal.pone.0269802. eCollection 2022.

ABSTRACT

OBJECTIVE: Stigma towards people with mental health problems is a growing issue across the world, to which healthcare providers might contribute. The aim of the present study was to explore psychiatrists’ attitudes towards their patients and link them to psychosocial and professional factors.

METHODS: An online questionnaire was used to approach the in- and outpatient psychiatric services across Hungary. A total of 211 trainees and specialists in adult and child psychiatry participated in our study. Their overall stigmatizing attitudes were measured, with focus on attitude, disclosure and help-seeking, and social distance dimensions by using the self-report Opening Minds Stigma Scale for Health Care Providers (OMS-HC). Multiple linear regression analyses were performed to elucidate the dimensions of stigma and its association with sociodemographic, professional and personal traits.

RESULTS: Stigmatizing attitudes of close colleagues towards patients were statistically significant predictors of higher scores on the attitude [B = 0.235 (0.168-0.858), p = 0.004], the disclosure and help-seeking subscales [B = 0.169 (0.038-0.908), p = 0.033], and the total score of the OMS-HC [B = 0.191 (0.188-1.843), p = 0.016]. Psychiatrists who had already sought help for their own problems had lower scores on the disclosure and help-seeking subscale [B = 0.202 (0.248-1.925), p = 0.011]. The overall stigmatizing attitude was predicted by the openness to participate in case discussion, supervision or Balint groups [B = 0.166 (0.178-5.886), p = 0.037] besides the more favorable attitudes of their psychiatrist colleagues [B = 0.191 (0.188-1.843), p = 0.016].

CONCLUSIONS: The favorable attitudes of psychiatrists are associated with their own experiences with any kind of psychiatric condition, previous help-seeking behavior and the opportunity to work together with fellow psychiatrists, whose attitudes are less stigmatizing. The perception of fellow colleagues’ attitudes towards patients and the openness to case discussion, supervision and Balint groups were the main two factors that affected the overall attitudes towards patients; therefore, these should be considered when tailoring anti-stigma interventions for psychiatrists.

PMID:35687584 | DOI:10.1371/journal.pone.0269802

PLoS Negl Trop Dis. 2022 Jun 10;16(6):e0010469. doi: 10.1371/journal.pntd.0010469. Online ahead of print.

ABSTRACT

Domestic dogs (Canis lupus familiaris) can transmit a variety of pathogens due to their ubiquitousness in urban, rural and natural environments, and their close interactions with wildlife and humans. In this study, we used a mixed-methods approach to assess the role of domestic dogs as potential intermediaries of disease transmission from wildlife to humans among indigenous Waiwai in the Konashen Community Owned Conservation Area, Guyana. To address these objectives we 1) performed physical examinations and collected biological samples to assess Waiwai domestic dog health, and 2) administered questionnaires to characterize the role of dogs in the community and identify potential transmission pathways between wildlife, dogs, and humans. We observed ectoparasites on all dogs (n = 20), including: fleas (100%), ticks (15%), botflies (30%), and jigger flea lesions (Tunga penetrans) (80%). Ten percent of dogs were seropositive for Ehrlichia canis/ewingii, 10% were positive for Dirofilaria immitis, and one dog was seropositive for Leishmania infantum. All dogs (n = 20) were seronegative for: canine distemper virus, Brucella canis, Leptospira serovars, Trypanosoma cruzi, Anaplasma phagocytophilum/platys and Borrelia burgdorferi. Our questionnaire data revealed that the Waiwai remove ectoparasites from their dogs, clean up dog feces, and administer traditional and/or Western medicine to their dogs. White blood cell, strongyle-type ova, and eosinophil counts were lower in dogs that were not frequently used for hunting, dogs that did receive traditional and/or western medicine, and dogs that were frequently kept in elevated dog houses, although differences were not statistically significant. While our results suggest that the Waiwai have developed cultural practices that may promote dog health and/or prevent zoonotic disease transmission, more research is necessary to determine the efficacy of these practices. Our study provides important data on the health of dogs and the potential for disease transmission to humans in a zoonotic hotspot.

PMID:35687596 | DOI:10.1371/journal.pntd.0010469

PLoS One. 2022 Jun 10;17(6):e0268600. doi: 10.1371/journal.pone.0268600. eCollection 2022.

ABSTRACT

BACKGROUND: Levels of child undernutrition and its correlates exhibit considerable spatial variation at different levels of granularity. In India, such variations and their interrelation have not been studied at the sub-district level primarily due to the non-availability of good quality granular data. Given the sheer regional diversity in India, it is essential to develop a region-specific evidence base at the micro-level.

DATA AND OBJECTIVES: The current study utilised, for the first time, a sub-district level survey data (Concurrent Child Monitoring Survey-II, 2014-15) to investigate the statistically significant clusters and spatial patterns of burden of undernutrition among children. The emergence of distinct patterns at the level of natural geographical regions of the state-coastal, southern and northern regions, lead to a region-specific analysis to measure the impact of various demographic, socio-economic and maternal factors on the prevalence of undernutrition specific to the three regions, using the National Family Health Survey-IV unit-level data.

METHODS: The spatial dependence and clustering of child undernourishment across sub-districts in Odisha were studied using various spatial statistical techniques, including spatial econometric models. Binary logistic regression was applied in the region-specific analysis.

RESULTS: Findings indicated statistically significant spatial clustering of undernutrition among children in specific geographic pockets with poor sanitation, low institutional and skilled deliveries, poor maternal health reinforcing the need for inter-sectoral coordination. Disparities across the three natural-regions, suggest that the parameters requiring priority for intervention may differ across levels of overall development.

CONCLUSION: The spatial clustering of different socio-demographic indicators in specific geographic pockets highlights the differential impact of these determinants on child undernutrition thereby reinforcing a strong need for targeted intervention in these areas. Present analysis and the evidence-based micro-level analysis can be utilised as a model for other Indian states and low-resource countries, making interventions more effective through multiple, synergistic and a multi-sectoral approach.

PMID:35687570 | DOI:10.1371/journal.pone.0268600

PLoS One. 2022 Jun 10;17(6):e0268586. doi: 10.1371/journal.pone.0268586. eCollection 2022.

ABSTRACT

Contact tracing is a key component of successful management of COVID-19. Contacts of infected individuals are asked to quarantine, which can significantly slow down (or prevent) community spread. Contact tracing is particularly effective when infections are detected quickly, when contacts are traced with high probability, when the initial number of cases is low, and when social distancing and border restrictions are in place. However, the magnitude of the individual contribution of these factors in reducing epidemic spread and the impact of population immunity (due to either previous infection or vaccination), in determining contact tracing outputs is not fully understood. We present a delayed differential equation model to investigate how the immunity status and the relaxation of social distancing requirements affect contact tracing practices. We investigate how the minimal contact tracing efficiency required to keep an outbreak under control depends on the contact rate and on the proportion of immune individuals. Additionally, we consider how delays in outbreak detection and increased case importation rates affect the number of contacts to be traced daily. We show that in communities that have reached a certain immunity status, a lower contact tracing efficiency is required to avoid a major outbreak, and delayed outbreak detection and relaxation of border restrictions do not lead to a significantly higher risk of overwhelming contact tracing. We find that investing in testing programs, rather than increasing the contact tracing capacity, has a larger impact in determining whether an outbreak will be controllable. This is because early detection activates contact tracing, which will slow, and eventually reverse exponential growth, while the contact tracing capacity is a threshold that will easily become overwhelmed if exponential growth is not curbed. Finally, we evaluate quarantine effectiveness in relation to the immunity status of the population and for different viral variants. We show that quarantine effectiveness decreases with increasing proportion of immune individuals, and increases in the presence of more transmissible variants. These results suggest that a cost-effective approach is to establish different quarantine rules for immune and nonimmune individuals, where rules should depend on viral transmissibility after vaccination or infection. Altogether, our study provides quantitative information for contact tracing downsizing in vaccinated populations or in populations that have already experienced large community outbreaks, to guide COVID-19 exit strategies.

PMID:35687566 | DOI:10.1371/journal.pone.0268586

PLoS One. 2022 Jun 10;17(6):e0269455. doi: 10.1371/journal.pone.0269455. eCollection 2022.

ABSTRACT

Untreated pain after surgery leads to poor patient satisfaction, longer hospital length of stay, lower health-related quality of life, and non-compliance with rehabilitation regimens. The aim of this study is to characterize the structure of acute pain trajectories during the postsurgical hospitalization period and quantify their association with pain at 30-days and 1-year after surgery. This cohort study included 2106 adult (≥18 years) surgical patients who consented to participate in the SATISFY-SOS registry (February 1, 2015 to September 30, 2017). Patients were excluded if they did not undergo invasive surgeries, were classified as outpatients, failed to complete follow up assessments at 30-days and 1-year following surgery, had greater than 4-days of inpatient stay, and/or recorded fewer than four pain scores during their acute hospitalization period. The primary exposure was the acute postsurgical pain trajectories identified by a machine learning-based latent class approach using patient-reported pain scores. Clinically meaningful pain (≥3 on a 0-10 scale) at 30-days and 1-year after surgery were the primary and secondary outcomes, respectively. Of the study participants (N = 2106), 59% were female, 91% were non-Hispanic White, and the mean (SD) age was 62 (13) years; 41% of patients underwent orthopedic surgery and 88% received general anesthesia. Four acute pain trajectory clusters were identified. Pain trajectories were significantly associated with clinically meaningful pain at 30-days (p = 0.007), but not at 1-year (p = 0.79) after surgery using covariate-adjusted logistic regression models. Compared to Cluster 1, the other clusters had lower statistically significant odds of having pain at 30-days after surgery (Cluster 2: [OR = 0.67, 95%CI (0.51-0.89)]; Cluster 3:[OR = 0.74, 95%CI (0.56-0.99)]; Cluster 4:[OR = 0.46, 95%CI (0.26-0.82)], all p<0.05). Patients in Cluster 1 had the highest cumulative likelihood of pain and pain intensity during the latter half of their acute hospitalization period (48-96 hours), potentially contributing to the higher odds of pain during the 30-day postsurgical period. Early identification and management of high-risk pain trajectories can help in ascertaining appropriate pain management interventions. Such interventions can mitigate the occurrence of long-term disabilities associated with pain.

PMID:35687544 | DOI:10.1371/journal.pone.0269455

Plast Reconstr Surg. 2022 Jun 13. doi: 10.1097/PRS.0000000000009307. Online ahead of print.

ABSTRACT

BACKGROUND: Social media has altered the mechanisms by which published research is disseminated and accessed. The objective of this study was to measure the effect of promotion on research article dissemination, influence, and impact in Plastic and Reconstructive Surgery.

METHODS: All articles published in Plastic and Reconstructive Surgery from January 1, 2016-December 31, 2018 were obtained and reviewed to determine inclusion/exclusion and for the Altmetric Attention Score (AAS), citations, relative citation rate (RCR), and 16 unique promotional tags (journal club, editor’s pick, press release, patient safety, etc.) as indexed on the Plastic and Reconstructive Surgery website. 1,502 articles were included in the analysis. Statistical analysis was completed using descriptive statistics, Pearson’s correlations, and Student t-tests where appropriate with a predetermined level of significance of p≤0.05.

RESULTS: A total of 637 articles (42.4%) had a promotional tag, while 252 (16.8%) had multiple tags. Articles with promotional tags had higher AAS (30.35 vs 8.22; p<0.001), more citations (11.96 vs 8.47; p<0.001), and a higher RCR (2.97 vs 2.06; p<0.001) compared to articles without a tag. Articles with multiple tags had higher AAS (50.17 vs 17.39; p<0.001), more citations (15.78 vs 9.47; p<0.001), and a higher RCR (3.67 vs 2.51; p<0.001) compared to articles with only one tag. As the number of tags increased for an article, AAS (p<0.001), citation count (p<0.001), and RCR (p<0.001) likewise increased.

CONCLUSIONS: This analysis strongly suggests that promotion of research articles is associated with significantly wider dissemination, broader visibility, and more subsequent citations in the literature.

PMID:35687416 | DOI:10.1097/PRS.0000000000009307

Am J Clin Nutr. 2022 Jun 10:nqac159. doi: 10.1093/ajcn/nqac159. Online ahead of print.

ABSTRACT

BACKGROUND: The causal effects of moderate alcohol consumption on cardiovascular diseases (CVDs) are continuously debated, especially on coronary heart disease (CHD).

OBJECTIVES: We aimed to explore the causal associations of alcohol consumption with CVDs and all-cause mortality among Chinese males.

METHODS: A prospective cohort study was conducted in 40,386 Chinese males, with 17,676 being genotyped for the rs671 variant in the aldehyde dehydrogenase 2 (ALDH2) gene. Cox proportional hazards model was conducted to estimate the effects of self-reported alcohol consumption. Mendelian randomization (MR) analysis was performed to explore the causality using rs671 as an instrumental variable.

RESULTS: During the follow-up of 303,353 person-years, 2,406 incident CVD and 3,195 all-cause mortality were identified. J-shaped associations of self-reported alcohol consumption with incident CVD and all-cause mortality were observed, showing decreased risks for light (≤25 g/day) and moderate drinkers (25-≤60 g/day). However, MR analyses revealed a linear association of genetically predicted alcohol consumption with the incident CVD (Ptrend = 0.02), including both CHD (Ptrend = 0.03) and stroke (Ptrend = 0.02). The HRs (95% CIs) for incident CVD across increasing tertiles of genetically predicted alcohol consumption were 1 (reference), 1.18 (1.01, 1.38), and 1.22 (1.03, 1.46). After excluding heavy drinkers, the risk of incident CVD and all-cause mortality was increased by 27% and 20% per standard drink increment of genetically predicted alcohol consumption, respectively.

CONCLUSIONS: Our analyses extend the evidence of the harmful effect of alcohol consumption to total CVD (including CHD) and all-cause mortality, highlighting the potential health benefits of lowering alcohol consumption, even among light-to-moderate male drinkers.

PMID:35687413 | DOI:10.1093/ajcn/nqac159

Plast Reconstr Surg. 2022 Jun 10. doi: 10.1097/PRS.0000000000009293. Online ahead of print.

ABSTRACT

BACKGROUND: Research in plastic surgery often includes bilateral procedures. This gives rise to issues with clustered data. Clustering is when individual data points within a data set are internally related. However, many authors do not account for clustering within their data which can lead to incorrect statistical conclusions.

METHODS: In February 2020, we searched PubMed to investigate the prevalence of reporting issues with bilateral breast procedures in plastic surgery literature. The review focused on breast surgery, since it often involves bilateral procedures and therefore clustering. Based on the review, we developed guidelines for how to identify and address clustered data. The guidelines were modified by a multidisciplinary group consisting of a biostatistician with expertise in clustered data at the Section of Biostatistics, University of Copenhagen, and three medical doctors and PhDs with expertise in statistical analysis and scientific methodology from the Copenhagen University Hospital, Rigshospitalet.

RESULTS: A total of 113 studies were included in the review. Seventy-five studies (66%) contained clustered data, but only 8 studies (11%) took clustering into account in the statistical analysis. These results were used to develop the CLUDA (CLUstered DAta) reporting guidelines which consists of two sections: one to identify clustering and one for reporting and analyzing clustered data.

CONCLUSIONS: Clustered data is abundant in plastic surgery literature, and we propose using the CLUDA reporting guidelines to identify and report clustered data and to consult a biostatistician when designing the study.

PMID:35687407 | DOI:10.1097/PRS.0000000000009293