Tag: statistics

PLoS One. 2022 Aug 11;17(8):e0272861. doi: 10.1371/journal.pone.0272861. eCollection 2022.

ABSTRACT

Kmeans clustering algorithm is an iterative unsupervised learning algorithm that tries to partition the given dataset into k pre-defined distinct non-overlapping clusters where each data point belongs to only one group. However, its performance is affected by its sensitivity to the initial cluster centroids with the possibility of convergence into local optimum and specification of cluster number as the input parameter. Recently, the hybridization of metaheuristics algorithms with the K-Means algorithm has been explored to address these problems and effectively improve the algorithm’s performance. Nonetheless, most metaheuristics algorithms require rigorous parameter tunning to achieve an optimum result. This paper proposes a hybrid clustering method that combines the well-known symbiotic organisms search algorithm with K-Means using the SOS as a global search metaheuristic for generating the optimum initial cluster centroids for the K-Means. The SOS algorithm is more of a parameter-free metaheuristic with excellent search quality that only requires initialising a single control parameter. The performance of the proposed algorithm is investigated by comparing it with the classical SOS, classical K-means and other existing hybrids clustering algorithms on eleven (11) UCI Machine Learning Repository datasets and one artificial dataset. The results from the extensive computational experimentation show improved performance of the hybrid SOSK-Means for solving automatic clustering compared to the standard K-Means, symbiotic organisms search clustering methods and other hybrid clustering approaches.

PMID:35951672 | DOI:10.1371/journal.pone.0272861

PLoS One. 2022 Aug 11;17(8):e0267705. doi: 10.1371/journal.pone.0267705. eCollection 2022.

ABSTRACT

INTRODUCTION: Genomic research and neurobiobanking are expanding globally. Empirical evidence on the level of awareness and willingness to donate/share biological samples towards the expansion of neurobiobanking in sub-Saharan Africa is lacking.

AIMS: To ascertain the awareness, perspectives and predictors regarding biological sample donation, sharing and informed consent preferences among community members in Ghana and Nigeria.

METHODS: A questionnaire cross-sectional survey was conducted among randomly selected community members from seven communities in Ghana and Nigeria.

RESULTS: Of the 1015 respondents with mean age 39.3 years (SD 19.5), about a third had heard of blood donation (37.2%, M: 42.4%, F: 32.0%, p = 0.001) and a quarter were aware of blood sample storage for research (24.5%; M: 29.7%, F: 19.4%, p = 0.151). Two out of ten were willing to donate brain after death (18.8%, M: 22.6%, F: 15.0%, p<0.001). Main reasons for unwillingness to donate brain were; to go back to God complete (46.6%) and lack of knowledge related to brain donation (32.7%). Only a third of the participants were aware of informed consent (31.7%; M: 35.9%, F: 27.5%, p<0.001). Predictors of positive attitude towards biobanking and informed consent were being married, tertiary level education, student status, and belonging to select ethnic groups.

CONCLUSION: There is a greater need for research attention in the area of brain banking and informed consent. Improved context-sensitive public education on neurobiobanking and informed consent, in line with the sociocultural diversities, is recommended within the African sub region.

PMID:35951660 | DOI:10.1371/journal.pone.0267705

PLoS Comput Biol. 2022 Aug 11;18(8):e1010367. doi: 10.1371/journal.pcbi.1010367. Online ahead of print.

ABSTRACT

Predictive modeling of drug-induced gene expressions is a powerful tool for phenotype-based compound screening and drug repurposing. State-of-the-art machine learning methods use a small number of fixed cell lines as a surrogate for predicting actual expressions in a new cell type or tissue, although it is well known that drug responses depend on a cellular context. Thus, the existing approach has limitations when applied to personalized medicine, especially for many understudied diseases whose molecular profiles are dramatically different from those characterized in the training data. Besides the gene expression, dose-dependent cell viability is another important phenotype readout and is more informative than conventional summary statistics (e.g., IC50) for characterizing clinical drug efficacy and toxicity. However, few computational methods can reliably predict the dose-dependent cell viability. To address the challenges mentioned above, we designed a new deep learning model, MultiDCP, to predict cellular context-dependent gene expressions and cell viability on a specific dosage. The novelties of MultiDCP include a knowledge-driven gene expression profile transformer that enables context-specific phenotypic response predictions of novel cells or tissues, integration of multiple diverse labeled and unlabeled omics data, the joint training of the multiple prediction tasks, and a teacher-student training procedure that allows us to utilize unreliable data effectively. Comprehensive benchmark studies suggest that MultiDCP outperforms state-of-the-art methods with unseen cell lines that are dissimilar from the cell lines in the supervised training in terms of gene expressions. The predicted drug-induced gene expressions demonstrate a stronger predictive power than noisy experimental data for downstream tasks. Thus, MultiDCP is a useful tool for transcriptomics-based drug repurposing and compound screening that currently rely on noisy high-throughput experimental data. We applied MultiDCP to repurpose individualized drugs for Alzheimer’s disease in terms of efficacy and toxicity, suggesting that MultiDCP is a potentially powerful tool for personalized drug discovery.

PMID:35951653 | DOI:10.1371/journal.pcbi.1010367

PLoS One. 2022 Aug 11;17(8):e0272333. doi: 10.1371/journal.pone.0272333. eCollection 2022.

ABSTRACT

BACKGROUND/OBJECTIVE: To evaluate the development of intra- and post-operative retinal breaks after pars plana vitrectomy (PPV) for macular hole (MH) and/or vitreomacular traction (VMT).

SUBJECTS/METHODS: Medical records of patients who underwent PPV at Kellogg Eye Center between 1/1/2005-6/30/2018, were evaluated in three groups: group 1, MH/VMT (n = 136); group 2, epiretinal membrane (ERM) without VMT (n = 270); and group 3, diagnostic vitrectomy (DV) or vitreous opacities (n = 35). Statistical analyses were conducted using SAS.

RESULTS: 20.6% of patients with MH/VMT, 8.5% of patients with ERM, and 5.7% of patients with DV or vitreous opacities had either intra-operative or post-operative breaks. Indication of MH/VMT versus ERM was a significant predictor for this outcome (p = .0112). The incidence of retinal breaks was higher in operations using 23-gauge versus 25-gauge PPV (25.0% vs. 7.4%, p < .0001).

CONCLUSIONS: The presence of MH and/or VMT is a significant risk factor for retinal breaks from PPV, as is use of 23-gauge vitrectomy.

PMID:35951646 | DOI:10.1371/journal.pone.0272333

PLoS One. 2022 Aug 11;17(8):e0272216. doi: 10.1371/journal.pone.0272216. eCollection 2022.

ABSTRACT

BACKGROUND: COVID-19 is a viral disease caused by a new strain of corona virus. Currently, prognosis and risk stratification of COVID-19 patients is done by the disease’s clinical presentation. Therefore, identifying laboratory biomarkers for disease prognosis and risk stratification of COVID-19 patients is critical for prompt treatment. Therefore, the main objective of this study was to assess the risk stratification and prognostic value of basic coagulation parameters and factors associated with disease severity among COVID-19 patients at the Tibebe Ghion Specialized Hospital, COVID-19 treatment center, Northwest Ethiopia.

METHODS: A follow-up study was conducted among conveniently recruited COVID-19 patients attended from March to June 2021. Socio-demographic and clinical data were collected using a structured questionnaire and checklist, respectively. Prothrombin time (PT) and activated partial thromboplastin time (APTT) were analyzed by the HUMACLOT DUE PLUS® machine. Descriptive statistics were used to summarize the socio-demographic and clinical characteristics of study participants. Kruskal Wallis tests were used to compare the difference between parametric and non-parametric continuous variables, respectively. The area under the receiver operating characteristic curve (AUC) was used to evaluate the value of PT and APTT in the risk stratification and disease prognosis of COVID-19 patients. Ordinal logistic regression was used to identify the factors associated with disease severity and prognosis. A P-value < 0.05 was defined as statistically significant for all results.

RESULT: Baseline PT at a cut-off value ≥ 16.25 seconds differentiated severe COVID-19 patients from mild and moderate patients (AUC: 0.89, 95% CI: 0.83-0.95). PT also differentiated mild COVID-19 patients from moderate and severe patients at a cut-off value ≤ 15.35 seconds (AUC: 0.90, 95% CI: 0.84-0.96). Moreover, alcohol drinkers were a 3.52 times more likely chance of having severe disease than non-drinkers (95% CI: 1.41-8.81). A one-year increment in age also increased the odds of disease severity by 6% (95% CI: 3-9%). An increment of ≥ 0.65 seconds from the baseline PT predicted poor prognosis (AUC: 0.93, 0.87-0.99).

CONCLUSIONS AND RECOMMENDATIONS: Prolonged baseline PT was observed in severe COVID-19 patients. Prolonged baseline PT was also predicted to worsen prognosis. An increase from the baseline PT was associated with worsen prognosis. Therefore, PT can be used as a risk stratification and prognostic marker in COVID-19 patients.

PMID:35951632 | DOI:10.1371/journal.pone.0272216

PLoS One. 2022 Aug 11;17(8):e0272903. doi: 10.1371/journal.pone.0272903. eCollection 2022.

ABSTRACT

Mycoplasma hyorhinis is an emerging swine pathogen bacterium causing polyserositis and polyarthritis in weaners and finishers. The pathogen is distributed world-wide, generating significant economic losses. No commercially available vaccine is available in Europe. Therefore, besides improving the housing conditions for prevention, antimicrobial therapy of the diseased animals is the only option to control the infection. Our aim was to determine the minimal inhibitory concentrations (MIC) of ten antimicrobials potentially used against M. hyorhinis infection. The antibiotic susceptibility of 76 M. hyorhinis isolates from Belgium, Germany, Hungary, Italy and Poland collected between 2019 and 2021 was determined by broth micro-dilution method and mismatch amplification mutation assay (MAMA). Low concentrations of tiamulin (MIC90 0.312 μg/ml), doxycycline (MIC90 0.078 μg/ml), oxytetracycline (MIC90 0.25 μg/ml), florfenicol (MIC90 2 μg/ml) and moderate concentrations of enrofloxacin (MIC90 1.25 μg/ml) inhibited the growth of the isolates. For the tested macrolides and lincomycin, a bimodal MIC pattern was observed (MIC90 >64 μg/ml for lincomycin, tulathromycin, tylosin and tilmicosin and 5 μg/ml for tylvalosin). The results of the MAMA assay were in line with the conventional method with three exceptions. Based on our statistical analyses, significant differences in MIC values of tiamulin and doxycycline were observed between certain countries. Our results show various levels of antimicrobial susceptibility among M. hyorhinis isolates to the tested antibiotics. The data underline the importance of susceptibility monitoring on pan-European level and provides essential information for proper antibiotic choice in therapy.

PMID:35951622 | DOI:10.1371/journal.pone.0272903

Adv Clin Exp Med. 2022 Aug 11. doi: 10.17219/acem/151640. Online ahead of print.

ABSTRACT

BACKGROUND: The assessment of the quality of life (QoL) in hematology-oncology patients is extremely important. The disease and anti-cancer therapy can cause adverse effects, directly impacting the physical and mental condition of the patient and indirectly influencing their social and professional situation. Therefore, a properly performed QoL assessment should take into account all of these aspects. Moreover, QoL assessment has a prognostic value in regard to treatment success and prognosis; therefore, the improvement in the QoL is often one of the goals of therapy.

OBJECTIVES: To identify the changes in QoL during therapy in patients with lymphoproliferative neoplasms.

MATERIAL AND METHODS: Forty-six hematology-oncology patients participated in this prospective single-center study. Their QoL was analyzed at 2 time points (before and after the first-line treatment). For this purpose, the EORTC QLQ-C30 questionnaire was used. All statistical analyses were performed using the STATISTICA v. 13 software. A value of p < 0.05 was considered statistically significant.

RESULTS: The study included patients with multiple myeloma (MM; 47.8%), non-Hodgkin lymphoma (NHL; 28.3%) and chronic lymphocytic leukemia (CLL; 23.9%). After the first line of treatment, patients perceived their overall QoL as slightly better than before starting the treatment, with an average increase of 1.94. Statistically significant differences were observed in physical and emotional functioning as well as fatigue, pain, dyspnea, appetite, and constipation.

CONCLUSIONS: In patients with lymphoproliferative neoplasms, after the first-line treatment, an improvement in an overall QoL and level of functioning, as well as a reduction in the severity of symptoms were observed.

PMID:35951626 | DOI:10.17219/acem/151640

PLoS One. 2022 Aug 11;17(8):e0272906. doi: 10.1371/journal.pone.0272906. eCollection 2022.

ABSTRACT

BACKGROUND: Human immunodeficiency virus (HIV) remains a global health threat, especially in developing countries. The successful scale-up of antiretroviral therapy (ART) programs to address this threat is hindered by a high proportion of patient loss to follow-up (LTFU). LTFU is associated with poor viral suppression and increased mortality. It is particularly acute among adolescents, who face unique adherence challenges. Although LTFU is a critical obstacle on the continuum of care for adolescents, few regional-level studies report the proportion of LTFU among adolescents receiving ART. Therefore, a systematic review and meta-analysis were conducted to estimate the pooled LTFU in ART programs among adolescents living with HIV in sub-Saharan Africa (SSA).

METHODS: We searched five databases (PubMed, Embase (Elsevier), PsycINFO, CINAHL, and Scopus) for articles published between 2005 and 2020 and reference lists of included articles. The PRISMA guidelines for systematic reviews were followed. A standardised checklist to extract data was used. Descriptive summaries were presented using narrative tables and figures. Heterogeneity within the included studies was examined using the Cochrane Q test statistics and I2 test. Random effect models were used to estimate the pooled prevalence of LTFU among ALHIV. We used Stata version 16 statistical software for our analysis.

RESULTS: Twenty-nine eligible studies (n = 285,564) were included. An estimated 15.07% (95% CI: 11.07, 19.07) of ALHIV were LTFU. Older adolescents (15-19 years old) were 43% (AOR = 0.57, 95% CI: 0.37, 0.87) more likely to be LTFU than younger (10-14 years old) adolescents. We find an insignificant relationship between gender and LTFU (AOR = 0.95, 95% CI: 0.87, 1.03). A subgroup analysis found that regional differences in the proportion of adolescent LTFU were not statistically significant. The trend analysis indicates an increasing proportion of adolescent LTFU over time.

CONCLUSIONS AND RECOMMENDATIONS: The proportion of LTFU among HIV-positive adolescents in SSA seems higher than those reported in other regions. Older adolescents in the region are at an increased risk for LTFU than younger adolescents. These findings may help policymakers develop appropriate strategies to retain ALHIV in ART services. Such strategies could include community ART distribution points, appointment spacing, adherence clubs, continuous free access to ART, and community-based adherence support.

PMID:35951621 | DOI:10.1371/journal.pone.0272906

Prev Chronic Dis. 2022 Aug 11;19:E49. doi: 10.5888/pcd19.210454.

ABSTRACT

INTRODUCTION: The density of tobacco retailers varies by community characteristics such as poverty levels or racial and ethnic composition. However, few studies have investigated how specific types of tobacco retailers vary by community characteristics. Our objective was to assess how the types of tobacco retailers in Ohio varied by the characteristics of the communities in which they were located.

RESULTS: For all US Census tracts, convenience stores were the most common type of retailer selling tobacco. Yet, the prevalence of convenience stores was higher in high-poverty urban tracts than in low-poverty urban tracts. Discount stores were the second-most common type of tobacco retailer and were most prevalent in rural tracts and high-racial and ethnic minority urban tracts. Grocery stores, pharmacies, and vape or hookah shops typically had the highest prevalence in more advantaged tracts.

CONCLUSION: Our findings demonstrate that the distribution of specific retailer types varies by community characteristics. The distribution of these retailer types has implications for product availability and price, which may subsequently affect tobacco use and cessation. To create equitable outcomes, policies should focus on retailers such as convenience and discount stores, which are heavily located in communities experiencing tobacco-related health disparities.

PMID:35951439 | DOI:10.5888/pcd19.210454

Health Psychol. 2022 Aug 11. doi: 10.1037/hea0001211. Online ahead of print.

ABSTRACT

OBJECTIVE: During the COVID-19 pandemic, the U.S. public was encouraged to practice good hand hygiene, such as hand washing or the use of hand sanitizer. Young adults reported lower levels of hand hygiene compared to adults of other ages. The aim of the current study was to test the effectiveness of different messages to promote hand sanitizer use among young adults.

METHOD: Over a 6-week period, we examined whether 3 brief messages (gain-frame, static descriptive norms, dynamic descriptive norms), placed next to sanitizer dispensers in university residence halls, predicted dispenser use in comparison to dispensers with no sign. Amount of sanitizer usage was measured 3 times per week via the weight of dispenser units. We tracked and controlled for the number of positive COVID-19 cases in residence halls because we expected it might influence sanitizer usage.

RESULTS: Compared to no signage, dispensers with signs had 35% greater usage, with the static descriptive norms sign associated with greatest usage (46% compared to no sign), although differences did not reach conventional levels of significance. The strongest predictor of sanitizer use was a residence hall’s degree of COVID-19 risk based on the hall’s case positivity.

CONCLUSIONS: Dispensers with signs had higher use than those without signs, but this difference was not statistically significant. We conclude that compared to prior research, “nudges” such as evidence-based messaging may have had less of an effect on health behavior engagement due to methodological differences across studies or characteristics of the COVID-19 context. (PsycInfo Database Record (c) 2022 APA, all rights reserved).

PMID:35951403 | DOI:10.1037/hea0001211