Shlomi Hod, Debanuj Nayak, Jason R. Gantenberg
et al.
Epidemiologic studies of infectious diseases often rely on models of contact networks to capture the complex interactions that govern disease spread, and ongoing projects aim to vastly increase the scale at which such data can be collected. However, contact networks may include sensitive information, such as sexual relationships or drug use behavior. Protecting individual privacy while maintaining the scientific usefulness of the data is crucial. We propose a privacy-preserving pipeline for disease spread simulation studies based on a sensitive network that integrates differential privacy (DP) with statistical network models such as stochastic block models (SBMs) and exponential random graph models (ERGMs). Our pipeline comprises three steps: (1) compute network summary statistics using \emph{node-level} DP (which corresponds to protecting individuals' contributions); (2) fit a statistical model, like an ERGM, using these summaries, which allows generating synthetic networks reflecting the structure of the original network; and (3) simulate disease spread on the synthetic networks using an agent-based model. We evaluate the effectiveness of our approach using a simple Susceptible-Infected-Susceptible (SIS) disease model under multiple configurations. We compare both numerical results, such as simulated disease incidence and prevalence, as well as qualitative conclusions such as intervention effect size, on networks generated with and without differential privacy constraints. Our experiments are based on egocentric sexual network data from the ARTNet study (a survey about HIV-related behaviors). Our results show that the noise added for privacy is small relative to other sources of error (sampling and model misspecification). This suggests that, in principle, curators of such sensitive data can provide valuable epidemiologic insights while protecting privacy.
Large language models (LLMs) have demonstrated impressive capabilities in disease diagnosis. However, their effectiveness in identifying rarer diseases, which are inherently more challenging to diagnose, remains an open question. Rare disease performance is critical with the increasing use of LLMs in healthcare settings. This is especially true if a primary care physician needs to make a rarer prognosis from only a patient conversation so that they can take the appropriate next step. To that end, several clinical decision support systems are designed to support providers in rare disease identification. Yet their utility is limited due to their lack of knowledge of common disorders and difficulty of use. In this paper, we propose RareScale to combine the knowledge LLMs with expert systems. We use jointly use an expert system and LLM to simulate rare disease chats. This data is used to train a rare disease candidate predictor model. Candidates from this smaller model are then used as additional inputs to black-box LLM to make the final differential diagnosis. Thus, RareScale allows for a balance between rare and common diagnoses. We present results on over 575 rare diseases, beginning with Abdominal Actinomycosis and ending with Wilson's Disease. Our approach significantly improves the baseline performance of black-box LLMs by over 17% in Top-5 accuracy. We also find that our candidate generation performance is high (e.g. 88.8% on gpt-4o generated chats).
The training of deep learning models relies on a large amount of labeled data. However, the high cost of medical labeling seriously hinders the development of deep learning in the medical field. Our study proposes a general disease diagnosis approach based on Zero-Shot Learning. The Siamese neural network is used to find similar diseases for the target diseases, and the U-Net segmentation model is used to accurately segment the key lesions of the disease. Finally, based on the ResNet-Agglomerative clustering algorithm, a clustering model is trained on a large number of sample data of similar diseases to obtain a approximate diagnosis of the target disease. Zero-Shot Learning of the target disease is then successfully achieved. To evaluate the validity of the model, we validated our method on a dataset of ophthalmic diseases in CFP modality. The external dataset was used to test its performance, and the accuracy=0.8395, precision=0.8094, recall=0.8463, F1 Score=0.8274, AUC=0.9226, which exceeded the indexes of most Few-Shot Learning and One-Shot Learning models. It proves that our method has great potential and reference value in the medical field, where annotation data is usually scarce and expensive to obtain.
Christian Fleming, Kelsey Mills, Nicolas Cardenas
et al.
The U.S. launched the Secure Pork Supply (SPS) Plan for Continuity of Business, a voluntary program providing foreign animal disease (FAD) guidance and setting biosecurity standards to maintain business continuity amid FAD outbreaks. The role of biosecurity in disease prevention is well recognized, yet the U.S. swine industry lacks knowledge of individual farm biosecurity plans and the efficacy of existing measures. We describe a multi-sector initiative that formed the Rapid Access Biosecurity (RAB) app consortium with the swine industry, government, and academia. We (i) summarized 7,625 farms using RABapp, (ii) mapped U.S. commercial swine coverage and areas of limited biosecurity, and (iii) examined associations between biosecurity and occurrences of porcine reproductive and respiratory syndrome virus (PRRSV) and porcine epidemic diarrhea virus (PEDV). RABapp, used in 31 states, covers ~47% of U.S. commercial swine. Of 307 Agricultural Statistics Districts with swine, 78% (238) had <50% of those animals in RABapp. We used a mixed-effects logistic regression model, accounting for production company and farm type (breeding vs. non-breeding). Requiring footwear/clothing changes, having multiple carcass disposal locations, hosting other businesses, and greater distance to swine farms reduced infection odds. Rendering carcasses, manure pit storage or land application, multiple perimeter buffer areas, and a larger animal housing area increased risk. This study leveraged RABapp to assess U.S. swine farm biosecurity, revealing gaps in SPS plan adoption that create vulnerable regions. Some biosecurity practices (e.g., footwear changes) lowered PRRSV/PEDV risk, while certain disposal and manure practices increased it. Targeted biosecurity measures and broader RABapp adoption can bolster industry resilience against foreign animal diseases.
The presence of intestinal parasites influences equines' well-being and working performance. However, the scenario of parasitism in working horses in the lowlands of Nepal is yet to be explored. The present study aimed to reveal the prevalence and diversity of intestinal parasites (protozoa and helminths) and to list the zoonotic species in working horses in the lowlands of Nepal. Fresh fecal samples (N=102) from horses were collected at two locations (Chitwan and Birgunj) in the lowlands of Terai and were transferred to the research laboratory. Coproscopy was carried out via direct wet mount, formalin ethyl acetate (FEA) sedimentation, saturated salt flotation, and acid-fast staining techniques. Coproscopy revealed an overall prevalence rate of 90.2% (92/102) with 15 known diverse species of parasites (Protozoa: 5 and Helminths: 10) and an unknown coccidian, out of which eight possess zoonotic potential. The prevalence and diversity of intestinal parasites were higher in adult than in young animals (90.7%; 15 spp. vs. 88.9%; 11 spp.) The overall prevalence of helminths was double that of protozoa (89.2% vs. 43.1%). Furthermore, polyparasitism was much more prevalent than monoparasitism (85.3% vs 4.9%). Co-infection with two parasite species (37%) was higher in young horses. In comparison, triplet infection (34%) was higher in adults, and a maximum concurrency of up to six species of parasites at a time was recorded. Following it, the differences in the prevalence rate of parasites based on the predictor of risks, like sex, grazing, domestication type, nature of the floor, and medication practices, were statistically significant. Working horses in the lowlands of Terai harbored a significant variety of intestinal parasites with important prevalence. Since eight of the reported parasitic species were zoonotic, infected horses pose a zoonotic risk to the owners. Therefore, timely deworming, pasture management, and reduction in working pressure are highly recommended.
Potato plants are plants that are beneficial to humans. Like other plants in general, potato plants also have diseases; if this disease is not treated immediately, there will be a significant decrease in food production. Therefore, it is necessary to detect diseases quickly and precisely so that disease control can be carried out effectively and efficiently. Classification of potato leaf disease can be done directly. Still, the symptoms cannot always explain the type of disease that attacks potato leaves because there are many types of diseases with symptoms that look the same. Humans also have deficiencies in determining the results of identification of potato leaf disease, so sometimes the results of identification between individuals can be different. Therefore, the use of Deep Learning for the classification process of potato leaf disease is expected to shorten the time and have a high classification accuracy. This study uses a deep learning method with the DenseNet201 architecture. The choice to use the DenseNet201 algorithm in this study is because the model can identify important features of potato leaves and recognize early signs of emerging diseases. This study aimed to evaluate the effectiveness of the transfer learning method with the DenseNet201 architecture in increasing the classification accuracy of potato leaf disease compared to traditional classification methods. This study uses two types of scenarios, namely, comparing the number of dropouts and comparing the three optimizers. This test produces the best model using dropout 0.1 and Adam optimizer with an accuracy of 99.5% for training, 95.2% for validation, and 96% for the confusion matrix. In this study, using data testing, as many as 40 images were tested into the model that has been built. The test results on this model resulted in a new accuracy for classifying potato leaf disease, namely 92.5%.
Theodore Pak, Sanjat Kanjilal, Cara McKenna
et al.
Background: Many US hospitals have stopped universal masking and testing all patients on admission for SARS-CoV-2. We assessed the association of ending universal masking and admission testing with the incidence of hospital-onset SARS-CoV-2 infections in five Massachusetts hospitals. Method: We conducted a retrospective study of all patients admitted between March 6, 2020 and December 14, 2023 and identified hospital-onset SARS-CoV-2 infections (newly positive SARS-CoV-2 PCR tests >4d after arrival) and community-onset infections (newly positive ≤4d after arrival). We excluded cases if local infection control teams discontinued precautions within 4d (suggesting a false positive or remote/resolved infection). We calculated weekly ratios between hospital-onset and community-onset SARS-CoV-2 cases to account for changes in community SARS-CoV-2 incidence over time. We then performed interrupted time series analysis, looking for changes in the ratio of hospital-onset to community-onset cases across three periods: pre-Omicron period with universal testing and masking in place (March 6, 2020–Dec 16, 2021); Omicron period with universal testing and masking in place (Dec 17, 2021–May 11, 2023); and Omicron period without universal testing and masking (May 12, 2023–Dec 14, 2023). We performed medical record reviews on 100 randomly selected hospital-onset cases after May 12, 2023 to examine if community-onset cases were being misclassified as hospital-onset cases. Result: During the study period, there were 626,908 patient admissions, including 24,980 with community-onset and 1,510 with hospital-onset SARS-CoV-2 infections. The mean weekly ratio of new hospital-onset to community-onset SARS-CoV-2 infections rose from 2.6% before Omicron, to 8.5% (95% CI, 7.0–9.9%) during Omicron, to 17% (95% CI, 15–19%) after universal admission testing and masking ended (Figure 1). There was a significant immediate level change after the pre-Omicron-to-Omicron transition (140% relative increase; 95% CI, 40–240%) and after universal admission testing and masking ended (110% relative increase; 95% CI, 73–150%). On medical record review of 100 randomly selected hospital-onset SARS-CoV-2 cases after universal admission testing had ended, 89% had new symptoms at the time of testing, 80% had PCR cycle thresholds ≤30, 27% had a known COVID-19 exposure, and 97% met at least one of these criteria. In-hospital mortality occurred in 8% of the 100 reviewed cases. Conclusion: Stopping universal masking and admission testing of all hospitalized patients at five Massachusetts hospitals was associated with a significant increase in hospital-onset COVID-19. Nosocomial COVID-19 remains a common complication of hospital care. Preventing nosocomial infections in this vulnerable population remains an important safety goal.
Infectious and parasitic diseases, Public aspects of medicine
We investigate optimal carbon abatement in a dynamic general equilibrium climate-economy model with endogenous structural change. By differentiating the production of investment from consumption, we show that social cost of carbon can be conceived as a reduction in physical capital. In addition, we distinguish two final sectors in terms of productivity growth and climate vulnerability. We theoretically show that heterogeneous climate vulnerability results in a climate-induced version of Baumol's cost disease. Further, if climate-vulnerable sectors have high (low) productivity growth, climate impact can either ameliorate (aggravate) the Baumol's cost disease, call for less (more) stringent climate policy. We conclude that carbon abatement should not only factor in unpriced climate capital, but also be tailored to Baumol's cost and climate diseases.
This work presents a model combining the simplest communicable and non-communicable disease models. The latter is, by far, the leading cause of sickness and death in the World, and introduces basal heterogeneity in populations where communicable diseases evolve. The model can be interpreted as a risk-structured model, another way of accounting for population heterogeneity. Our results show that considering the non-communicable disease (in the end, heterogeneous populations) allows the communicable disease to become endemic even if the basic reproduction number is less than $1$. This feature is known as subcritical bifurcation. Furthermore, ignoring the non-communicable disease dynamics results in overestimating the reproduction number and, thus, giving wrong information about the actual number of infected individuals. We calculate sensitivity indices and derive interesting epidemic-control information.
There is rising evidence of the health benefit associated with specific dietary interventions. Current food-disease databases focus on associations and treatment relationships but haven't provided a reasonable assessment of the strength of the relationship, and lack of attention on food nutrition. There is an unmet need for a large database that can guide dietary therapy. We fill the gap with NutriFD, a scoring network based on associations and therapeutic relationships between foods and diseases. NutriFD integrates 9 databases including foods, nutrients, diseases, genes, miRNAs, compounds, disease ontology and their relationships. To our best knowledge, this database is the only one that can score the associations and therapeutic relationships of everyday foods and diseases by weighting inference scores of food compounds to diseases. In addition, NutriFD demonstrates the predictive nature of nutrients on the therapeutic relationships between foods and diseases through machine learning models, laying the foundation for a mechanistic understanding of food therapy.
Intro: Dengue virus (DENV) is one of the most important arboviral human pathogens with an estimated 25,000 deaths/year worldwide. Biological control of Aedes-borne transmission using the endosymbiotic bacteria Wolbachia has resulted in a reduction of dengue incidence in endemic areas. However, there is a concern that DENV, an RNA virus, could mutate to escape Wolbachia blocking, leading to reduced effectiveness of this intervention. Little is known about the intracellular interactions between the virus and Wolbachia during coinfection of mosquitoes, and any Wolbachia-mediated effect on DENV genetic diversity. Methods: Two DENV strains were used, and infections performed on paired C6/36 mosquito cell lines, one infected with the Wolbachia strain wAlbB and an uninfected, control cell line. Supernatants and cells were sampled at several time points over a period of 6 days. Resulting virus titres were evaluated using qRT-PCR and DENV genetic diversity throughout infection determined using nanopore long-read sequencing. Intracellular localisation between Wolbachia and DENV at several timepoints was determined using transmission electron microscopy (TEM). Findings: Changes in DENV genetic diversity were compared between serotypes and any amino acid changes observed were modelled to determine their functional impact on the virus. We determined whether Wolbachia and DENV colocalise to the endoplasmic reticulum and investigated any subcellular changes during the time course of infection. Cellular structural changes were correlated to changes in viral titre and viral genetic diversity. Discussion: Sampling and evaluation of viral titre, selection pressure, and intracellular interactions throughout the time-course of infection will provide a broad picture of how Wolbachia and DENV interact within the mosquito cell. Understanding the long-term ability of Wolbachia to continue blocking DENV transmission relies on a thorough understanding of underlying cellular interactions and mechanisms. Conclusion: The results shed light on DENV localisation within Wolbachia- infected cells and the fine scale genetic changes in the virus that may occur during infection.
Vitor Cardoso, Taishi Ikeda, Rodrigo Vicente
et al.
Fuzzy dark matter is an exciting alternative to the standard cold dark matter paradigm, reproducing its large scale predictions, while solving most of the existing tension with small scale observations. These models postulate that dark matter is constituted by light bosons and predict the condensation of a solitonic core -- also known as boson star, supported by wave pressure -- at the center of halos. However, solitons which host a \emph{parasitic} supermassive black hole are doomed to be swallowed by their guest. It is thus crucial to understand in detail the accretion process. In this work, we use numerical relativity to self-consistently solve the problem of accretion of a boson star by a central black hole, in spherical symmetry. We identify three stages in the process, a {\it boson-quake}, a {\it catastrophic stage} and a linear phase, as well as a general accurate expression for the lifetime of a boson star with an endoparasitic black hole. Lifetimes of these objects can be large enough to allow them to survive until the present time.
Maduja VM Divarathna, Rukshan AM Rafeek, Sampath Jayaweera
et al.
Respiratory Syncytial Virus (RSV) is one of the most common respiratory viruses causing acute respiratory tract infections (ARTI) in children. Detailed data on RSV infections including the RSV types circulating in Sri Lanka are not available. This study aimed to determine the prevalence, patterns and characterization of RSV associated ARTI in hospitalized children less than 5 years in a general hospital in Sri Lanka. We tested 500 nasopharyngeal aspirate (NPA) samples collected from children with suspected viral ARTI from May 2016 to July 2018 from Kegalle General Hospital, Sri Lanka for RSV using antigen detection by an immunofluorescence assay (IFA). RSV positive samples were further characterized using the real time RT-PCR. RSV was the predominant virus associated with ARTI with a prevalence of 28% (140/500) in the study sample. RSV in was also detected in more co-infections with other respiratory viruses. RSV was detected throughout the year with peak periods from June to August 2016, March to July 2017 and May to July 2018. Of the 140 RSV positive children tested, 72.14% had RSV-B, while 27.86% had RSV-A infection. Both RSV subtypes were detected throughout the study period with overlapping patterns. A few co-infections between RSV-A and RSV-B were detected during the co-circulation. RSV was the most prevalent virus and RSV-B was the predominant subgroup associated with ARTI in the children <5 years in Sri Lanka from May 2016 to July 2018. RSV was detected throughout the study period with peaks in certain months in the study area.
Abstract Background Schistosoma japonicum infection is an important public health problem, imposing heavy social and economic burdens in 78 countries worldwide. However, the mechanism of transition from chronic to advanced S. japonicum infection remains largely unknown. Evidences suggested that gut microbiota plays a role in the pathogenesis of S. japonicum infection. However, the composition of the gut microbiota in patients with chronic and advanced S. japonicum infection is not well defined. In this study, we compared the composition of the intestinal flora in patients with chronic and advanced S. japonicum infection. Methods The feces of 24 patients with chronic S. japonicum infection and five patients with advanced S. japonicum infection from the same area were collected according to standard procedures, and 16S rRNA sequencing technology was used to analyze the intestinal microbial composition of the two groups of patients. Results We found that alteration occurs in the gut microbiota between the groups of patients with chronic and advanced S. japonicum infections. Analysis of alpha and beta diversity indicated that the diversity and abundance of intestinal flora in patients with advanced S. japonicum infection were lower than those in patients with chronic S. japonicum infection. Furthermore, Prevotella 9, Subdoligranulum, Ruminococcus torques, Megamonas and Fusicatenibacter seemed to have potential to discriminate different stages of S. japonicum infection and to act as biomarkers for diagnosis. Function prediction analysis revealed that microbiota function in the chronic group was focused on translation and cell growth and death, while that in the advanced group was concentrated on elevating metabolism-related functions. Conclusions Our study demonstrated that alteration in gut microbiota in different stages of S. japonicum infection plays a potential role in the pathogenesis of transition from chronic to advanced S. japonicum infection. However, further validation in the clinic is needed, and the underlying mechanism requires further study.
We focus on the modeling and simulation of an infectious disease spreading in a medium size population occupying a confined environment, such as an airport terminal, for short periods of time. Because of the size of the crowd and venue, we opt for a kinetic type model. The paper is divided into two parts. In the first part, we adopt the simplifying assumption that people's walking speed and direction are given. The resulting kinetic model features a variable that denotes the level of exposure to people spreading the disease, a parameter describing the contagion interaction strength, and a kernel function that is a decreasing function of the distance between a person and a spreading individual. Such model is tested on problems involving a small crowd in a square walkable domain. In the second part, ideas from the simplified model are used to incorporate disease spreading in a kinetic theory approach for crowd dynamics, i.e. the walking speed and direction result from interaction with other people and the venue.