Category: Statistics

PLoS Negl Trop Dis. 2023 Feb 8;17(2):e0011067. doi: 10.1371/journal.pntd.0011067. eCollection 2023 Feb.

ABSTRACT

BACKGROUND: Melioidosis is a disease caused by the bacterium Burkholderia pseudomallei, infecting humans and non-human primates (NHP) through contaminated soil or water. World-wide there are an estimated 165,000 human melioidosis cases each year, but recordings of NHP cases are sporadic. Clinical detection of melioidosis in humans is primarily by culturing B. pseudomallei, and there are no standardized detection protocols for NHP. NHP are an important animal model for melioidosis research including clinical trials and development of biodefense countermeasures.

METHODOLOGY/PRINCIPLE FINDINGS: We evaluated the diagnostic potential of the multiple antigen serological assay, BurkPx, in NHP using two sera sets: (i) 115 B. pseudomallei-challenged serum samples from 80 NHP collected each week post-exposure (n = 52) and at euthanasia (n = 47), and (ii) 126 B. pseudomallei-naïve/negative serum samples. We observed early IgM antibody responses to carbohydrate antigens followed by IgG antibody recognition to multiple B. pseudomallei protein antigens during the second week of infection. B. pseudomallei negative serum samples had low to intermediate antibody cross reactivity to the antigens in this assay. Infection time was predicted as the determining factor in the variation of antibody responses, with 77.67% of variation explained by the first component of the principal component analysis. A multiple antigen model generated a binary prediction metric ([Formula: see text]), which when applied to all data resulted in 100% specificity and 63.48% sensitivity. Removal of week 1 B. pseudomallei challenged serum samples increased the sensitivity of the model to 95%.

CONCLUSION/SIGNIFICANCE: We employed a previously standardized assay for humans, the BurkPx assay, and assessed its diagnostic potential for detection of B. pseudomallei exposure in NHP. The assay is expected to be useful for surveillance in NHP colonies, in investigations of suspected accidental releases or exposures, and for identifying vaccine correlates of protection.

PMID:36753522 | DOI:10.1371/journal.pntd.0011067

PLoS Comput Biol. 2023 Feb 8;19(2):e1010878. doi: 10.1371/journal.pcbi.1010878. Online ahead of print.

ABSTRACT

Humans constantly assess the appearance of materials to plan actions, such as stepping on icy roads without slipping. Visual inference of materials is important but challenging because a given material can appear dramatically different in various scenes. This problem especially stands out for translucent materials, whose appearance strongly depends on lighting, geometry, and viewpoint. Despite this, humans can still distinguish between different materials, and it remains unsolved how to systematically discover visual features pertinent to material inference from natural images. Here, we develop an unsupervised style-based image generation model to identify perceptually relevant dimensions for translucent material appearances from photographs. We find our model, with its layer-wise latent representation, can synthesize images of diverse and realistic materials. Importantly, without supervision, human-understandable scene attributes, including the object’s shape, material, and body color, spontaneously emerge in the model’s layer-wise latent space in a scale-specific manner. By embedding an image into the learned latent space, we can manipulate specific layers’ latent code to modify the appearance of the object in the image. Specifically, we find that manipulation on the early-layers (coarse spatial scale) transforms the object’s shape, while manipulation on the later-layers (fine spatial scale) modifies its body color. The middle-layers of the latent space selectively encode translucency features and manipulation of such layers coherently modifies the translucency appearance, without changing the object’s shape or body color. Moreover, we find the middle-layers of the latent space can successfully predict human translucency ratings, suggesting that translucent impressions are established in mid-to-low spatial scale features. This layer-wise latent representation allows us to systematically discover perceptually relevant image features for human translucency perception. Together, our findings reveal that learning the scale-specific statistical structure of natural images might be crucial for humans to efficiently represent material properties across contexts.

PMID:36753520 | DOI:10.1371/journal.pcbi.1010878

PLoS One. 2023 Feb 8;18(2):e0280623. doi: 10.1371/journal.pone.0280623. eCollection 2023.

ABSTRACT

A major limiting factor in combatting the HIV epidemic has been the identification of people living with HIV. Index testing programs were developed to face that challenge. Index testing is a focused HIV testing service approach in which family members and partners of people living with HIV are offered testing. Despite the implementation of index testing, there is still a gap between the estimated number of people living with HIV and those who know their status in Côte d’Ivoire. This study aimed to understand the implementation process of index testing in Côte d’Ivoire and to identify implementation challenges from healthcare workers perspectives. In January and February 2020, we conducted a qualitative study through 105 individual semi-structured interviews regarding index testing with clinical providers (physicians, nurses, and midwives) and non-clinical providers (community counselors and their supervisors) at 16 rural health facilities across four regions of Côte d’Ivoire. We asked questions regarding the index testing process, index client intake, contact tracing and testing, the challenges of implementation, and solicited recommendations on improving index testing in Côte d’Ivoire. The interviews revealed that index testing is implemented by non-clinical providers. Passive referral, by which the index client brought their contact to be tested, and providers referral, by which a healthcare worker reached out to the index client’s contact, were the preferred contact tracing and testing strategies. There was not statistically significant difference between immediate and delayed notification. Reported challenges of index testing implementation included index cases refusing to give their partner’s information or a partner refusing to be tested, fear of divorce, societal stigma, long distances, lack of appropriate training in index testing strategies, and lack of a private room for counseling. The recommendations given by providers to combat these was to reinforce HIV education among the population, to train healthcare workers on index testing strategies, and to improve infrastructure, transportation, and communication resources. The study showed that the elements that influenced the process of index testing in Côte d’Ivoire were multifactorial, including individual, interpersonal, health systems, and societal factors. Thus, a multi-faceted approach to overcoming challenges of index testing in Côte d’Ivoire is needed to improve the yield of index testing.

PMID:36753518 | DOI:10.1371/journal.pone.0280623

Inflamm Bowel Dis. 2023 Feb 8:izad013. doi: 10.1093/ibd/izad013. Online ahead of print.

ABSTRACT

BACKGROUND: High histologic remission rates have been reported with placebos in randomized controlled trials (RCTs) evaluating ulcerative colitis (UC) therapies and have varied based on trial designs. We performed a systematic review and meta-analysis to quantify placebo histological remission rates and identify factors influencing those rates.

METHODS: MEDLINE, EMBASE, and the Cochrane library were searched from inception of the databases until December 2021. We included placebo-controlled RCTs of adult patients with UC treated with aminosalicylates, corticosteroids, immunosuppressives, biologics, and small molecules. We pooled estimates using a random-effects model and performed subgroup analysis and meta-regression to evaluate the effect of different covariates on placebo rates.

RESULTS: Thirty-three studies (30 induction and 3 maintenance) were included. The overall placebo histological remission rate was 15.7% (95% confidence interval, 12.9%-19%) across all 33 studies. High heterogeneity was observed among studies with I2 = 62.10%. The pooled estimate of histological remission was 15.8% in induction studies and 14.5% in maintenance studies. Subgroup analysis revealed statistically significant differences in placebo rates when accounting for background medications, the intervention drug class, and disease severity (P = .041, .025, and .025, respectively). There was no statistical difference between induction vs maintenance studies or between different histological scales (P = .771, and .075, respectively).

CONCLUSIONS: Placebo histological remission rates range from 13% to 19% in UC RCTs, but studies are highly heterogeneous. Factors found to influence placebo rates include presence of background medications, the drug used, and the disease severity. These observations inform future trial designs to minimize placebo rates and reduce heterogeneity.

PMID:36753516 | DOI:10.1093/ibd/izad013

PLoS Negl Trop Dis. 2023 Feb 8;17(2):e0011072. doi: 10.1371/journal.pntd.0011072. eCollection 2023 Feb.

ABSTRACT

Burkholderia pseudomallei, the causative agent of melioidosis, is a gram-negative soil bacterium well recognized in Southeast Asia and northern Australia. However, wider and expanding global distribution of B. pseudomallei has been elucidated. Early diagnosis is critical for commencing the specific therapy required to optimize outcome. Serological testing using the indirect hemagglutination (IHA) antibody assay has long been used to augment diagnosis of melioidosis and to monitor progress. However, cross reactivity and prior exposure may complicate the diagnosis of current clinical disease (melioidosis). The goal of our study was to develop and initially evaluate a serology assay (BurkPx) that capitalized upon host response to multiple antigens. Antigens were selected from previous studies for expression/purification and conjugation to microspheres for multiantigen analysis. Selected serum samples from non-melioidosis controls and serial samples from culture-confirmed melioidosis patients were used to characterize the diagnostic power of individual and combined antigens at two times post admission. Multiple variable models were developed to evaluate multivariate antigen reactivity, identify important antigens, and determine sensitivity and specificity for the diagnosis of melioidosis. The final multiplex assay had a diagnostic sensitivity of 90% and specificity of 93%, which was superior to any single antigen in side-by-side comparisons. The sensitivity of the assay started at >85% for the initial serum sample after admission and increased to 94% 21 days later. Weighting antigen contribution to each model indicated that certain antigen contributed to diagnosis more than others, which suggests that the number of antigens in the assay can be decreased. In summation, the BurkPx assay can facilitate the diagnosis of melioidosis and potentially improve on currently available serology assays. Further evaluation is now required in both melioidosis-endemic and non-endemic settings.

PMID:36753506 | DOI:10.1371/journal.pntd.0011072

PLoS One. 2023 Feb 8;18(2):e0281474. doi: 10.1371/journal.pone.0281474. eCollection 2023.

ABSTRACT

In this paper, we introduced a novel general two-parameter statistical distribution which can be presented as a mix of both exponential and gamma distributions. Some statistical properties of the general model were derived mathematically. Many estimation methods studied the estimation of the proposed model parameters. A new statistical model was presented as a particular case of the general two-parameter model, which is used to study the performance of the different estimation methods with the randomly generated data sets. Finally, the COVID-19 data set was used to show the superiority of the particular case for fitting real-world data sets over other compared well-known models.

PMID:36753497 | DOI:10.1371/journal.pone.0281474

PLoS One. 2023 Feb 8;18(2):e0281466. doi: 10.1371/journal.pone.0281466. eCollection 2023.

ABSTRACT

BACKGROUND: Polypharmacy can be a consequence of overprescribing that is prevalent in older adults with multimorbidity. Polypharmacy can cause adverse reactions and result in hospital admission. This study predicted risks of adverse drug reaction (ADR)-related and emergency hospital admissions by medicine classes.

METHODS: We used electronic health record data from general practices of Clinical Practice Research Datalink (CPRD GOLD) and Aurum. Older patients who received at least five medicines were included. Medicines were classified using the British National Formulary sections. Hospital admission cases were propensity-matched to controls by age, sex, and propensity for specific diseases. The matched data were used to develop and validate random forest (RF) models to predict the risk of ADR-related and emergency hospital admissions. Shapley Additive eXplanation (SHAP) values were calculated to explain the predictions.

RESULTS: In total, 89,235 cases with polypharmacy and hospitalised with an ADR-related admission were matched to 443,497 controls. There were over 112,000 different combinations of the 50 medicine classes most implicated in ADR-related hospital admission in the RF models, with the most important medicine classes being loop diuretics, domperidone and/or metoclopramide, medicines for iron-deficiency anaemias and for hypoplastic/haemolytic/renal anaemias, and sulfonamides and/or trimethoprim. The RF models strongly predicted risks of ADR-related and emergency hospital admission. The observed Odds Ratio in the highest RF decile was 7.16 (95% CI 6.65-7.72) in the validation dataset. The C-statistics for ADR-related hospital admissions were 0.58 for age and sex and 0.66 for RF probabilities.

CONCLUSIONS: Polypharmacy involves a very large number of different combinations of medicines, with substantial differences in risks of ADR-related and emergency hospital admissions. Although the medicines may not be causally related to increased risks, RF model predictions may be useful in prioritising medication reviews. Simple tools based on few medicine classes may not be effective in identifying high risk patients.

PMID:36753492 | DOI:10.1371/journal.pone.0281466

PLoS One. 2023 Feb 8;18(2):e0276827. doi: 10.1371/journal.pone.0276827. eCollection 2023.

ABSTRACT

BACKGROUND: Mobile-linked point-of-care diagnostics forms an integral part of diagnostic health services for efficient communication between patients and healthcare professionals despite geographical location and time of diagnosis. The efficiency of this technology lies in the user experience which means that the interaction of the user with the implemented technology needs to be simple, convenient, and consistent. Having a well-structured user experience of these devices in community-based healthcare will aid in sustainable implementation. Herein, we propose to conduct a literature search to systematically map out evidence based on mobile-linked POC diagnostics user experience at a community level in resource-limited settings.

METHODOLOGY: The proposed scoping review will be guided by the advanced Arksey and O’Malley methodological framework and further advanced by Levac et al. A comprehensive search will be conducted to find relevant published literature from the following electronic databases: Scopus, Web of Science, EBSCOhost (Medline, CINAHL, Africa-wide, Academic Search Complete). Grey literature will also be searched, including reports from government and international organizations such as World Health Organization (WHO), Foundation for Innovative New Diagnostics (FIND), and the Food and Drug Administration (FDA). Two independent reviewers will screen the relevant studies and the degree of the agreement will be determined by calculating Cohen’s kappa statistic. The quality of eligible data will also be appraised using the mixed method appraisal tool version 2018.

DISCUSSION: We anticipate that the planned scoping review will present useful evidence to inform stakeholders on the integration of mobile-linked diagnostic devices in community-based healthcare which will guide further research on the subject.

PMID:36753489 | DOI:10.1371/journal.pone.0276827

PLoS One. 2023 Feb 8;18(2):e0278555. doi: 10.1371/journal.pone.0278555. eCollection 2023.

ABSTRACT

The Theory of Evolution (TE) is the backbone of biology and is the best way to explain the diversity of species that exist on the planet. However, despite all the supporting evidence, TE remains poorly understood and accepted. In this study, the levels of acceptance and understanding of TE were measured, respectively, using the Inventory of Student Evolution Acceptance (I-SEA) and Knowledge of Evolution Exam (KEE) questionnaires, in high school students in Monterrey, Mexico (N = 370). The results show that the acceptance of TE ranges from moderate (90.3 out of 120) to high (3.7 out of 5), depending on the scale with which it is measured, while the level of comprehension is low (4.5 out of 10). Statistical analysis of the data collected reveals that there is a positive relationship between acceptance and understanding of TE (r = 0.34). In addition, the proportions of I-SEA and KEE were evaluated based on several factors, such as religion and educational level of the parents, among others. It was found that the level of education of the parents positively affects the understanding of the basic concepts of TE, while religion is the main factor of negative influence on both acceptance and understanding. Finally, the low comprehension shown in this study suggests a revision and readjustment of the contents that are taught in the upper secondary education curriculum.

PMID:36753485 | DOI:10.1371/journal.pone.0278555

JMIR Med Educ. 2023 Feb 8;9:e45312. doi: 10.2196/45312.

ABSTRACT

BACKGROUND: Chat Generative Pre-trained Transformer (ChatGPT) is a 175-billion-parameter natural language processing model that can generate conversation-style responses to user input.

OBJECTIVE: This study aimed to evaluate the performance of ChatGPT on questions within the scope of the United States Medical Licensing Examination Step 1 and Step 2 exams, as well as to analyze responses for user interpretability.

METHODS: We used 2 sets of multiple-choice questions to evaluate ChatGPT’s performance, each with questions pertaining to Step 1 and Step 2. The first set was derived from AMBOSS, a commonly used question bank for medical students, which also provides statistics on question difficulty and the performance on an exam relative to the user base. The second set was the National Board of Medical Examiners (NBME) free 120 questions. ChatGPT’s performance was compared to 2 other large language models, GPT-3 and InstructGPT. The text output of each ChatGPT response was evaluated across 3 qualitative metrics: logical justification of the answer selected, presence of information internal to the question, and presence of information external to the question.

RESULTS: Of the 4 data sets, AMBOSS-Step1, AMBOSS-Step2, NBME-Free-Step1, and NBME-Free-Step2, ChatGPT achieved accuracies of 44% (44/100), 42% (42/100), 64.4% (56/87), and 57.8% (59/102), respectively. ChatGPT outperformed InstructGPT by 8.15% on average across all data sets, and GPT-3 performed similarly to random chance. The model demonstrated a significant decrease in performance as question difficulty increased (P=.01) within the AMBOSS-Step1 data set. We found that logical justification for ChatGPT’s answer selection was present in 100% of outputs of the NBME data sets. Internal information to the question was present in 96.8% (183/189) of all questions. The presence of information external to the question was 44.5% and 27% lower for incorrect answers relative to correct answers on the NBME-Free-Step1 (P<.001) and NBME-Free-Step2 (P=.001) data sets, respectively.

CONCLUSIONS: ChatGPT marks a significant improvement in natural language processing models on the tasks of medical question answering. By performing at a greater than 60% threshold on the NBME-Free-Step-1 data set, we show that the model achieves the equivalent of a passing score for a third-year medical student. Additionally, we highlight ChatGPT’s capacity to provide logic and informational context across the majority of answers. These facts taken together make a compelling case for the potential applications of ChatGPT as an interactive medical education tool to support learning.

PMID:36753318 | DOI:10.2196/45312