Mitigating Machine Learning Bias Between High Income and Low-Middle Income Countries for Enhanced Model Fairness and Generalizability (preprint)

Collaborative efforts in artificial intelligence (AI) are increasingly common between high-income countries (HICs) and low- to middle-income countries (LMICs). Given the resource limitations often encountered by LMICs, collaboration becomes crucial for pooling resources, expertise, and knowledge. Despite the apparent advantages, ensuring the fairness and equity of these collaborative models is essential, especially considering the distinct differences between LMIC and HIC hospitals. In this study, we show that collaborative AI approaches can lead to divergent performance outcomes across HIC and LMIC settings, particularly in the presence of data imbalances. Through a real-world COVID-19 screening case study, we demonstrate that implementing algorithmic-level bias mitigation methods significantly improves outcome fairness between HIC and LMIC sites while maintaining high diagnostic sensitivity. We compare our results against previous benchmarks, utilizing datasets from four independent United Kingdom Hospitals and one Vietnamese hospital, representing HIC and LMIC settings, respectively.

SARS-CoV-2 sequencing with cloud-based analysis illustrates expedient co-ordinated surveillance of viral genomic epidemiology across six continents (preprint)

Viral sequencing has been critical in the COVID-19 pandemic response, but sequencing and bioinformatics capacity remain inconsistent. To examine the utility of a cloud-based sequencing analysis platform for SARS-CoV-2 sequencing, we conducted a cross-sectional study incorporating seven countries in July 2022. Sites submitted sequential SARS-CoV-2 sequences over two weeks to the Global Pathogen Analysis Service (GPAS). The GPAS bioinformatics cloud platform performs sequence assembly plus lineage and related sample identification. Users can share information with collaborators while retaining data ownership. Seven sites contributed sequencing reads from 5,346 clinical samples, of which 4,799/5,346 (89.8%) had a lineage identified. Omicron lineages dominated, with the vast majority being BA.5, BA.4 and BA.2, commensurate with contemporary genomic epidemiological observations. Phylogenetic analysis demonstrated low within-lineage diversity, and highly similar sequences present in globally disparate sites. A cloud-based analysis platform like GPAS addresses bioinformatics bottlenecks and facilitates collaboration in pathogen surveillance, enhancing epidemic and pandemic preparedness.

Upgrading nirmatrelvir to inhibit SARS-CoV-2 Mpro via DeepFrag and free energy calculations.

The first oral drug for the treatment of COVID-19, Paxlovid, has been authorized; however, nirmatrelvir, a major component of the drug, is reported to be associated with some side effects. Moreover, the appearance of many novel variants raises concerns about drug resistance, and designing new potent inhibitors to prevent viral replication is thus urgent. In this context, using a hybrid approach combining machine learning (ML) and free energy simulations, 6 compounds obtained by modifying nirmatrelvir were proposed to bind strongly to SARS-CoV-2 Mpro. The structural modification of nirmatrelvir significantly enhances the electrostatic interaction free energy between the protein and ligand and slightly decreases the vdW term. However, the vdW term is the most important factor in controlling the ligand-binding affinity. In addition, the modified nirmatrelvir might be less toxic to the human body than the original inhibitor.

Changes in the care of acute cerebrovascular and cardiovascular conditions during the first year of the covid-19 pandemic in 746 hospitals in the USA: retrospective analysis.

Objective: To measure the impact of the covid-19 pandemic on admissions to hospital and interventions for acute ischemic stroke and acute myocardial infarction. Design: A retrospective analysis. Setting: 746 qualifying hospitals in the USA from the Premier Healthcare Database. Participants: Patients aged 18 years and older who were admitted to hospital with a primary diagnosis of acute ischemic stroke or acute myocardial infarction between 1 March 2019 and 28 February 2021. Main outcome measures: Relative changes in volumes were assessed for acute ischemic stroke and acute myocardial infarction hospital admissions as well as intravenous thrombolysis, mechanical thrombectomy, and percutaneous coronary intervention (overall and for acute myocardial infarction only) across the first year of the pandemic versus the prior year. Mortality in hospital and length of stay in hospital were also compared across the first year of the pandemic versus the corresponding period the year prior. These metrics were explored across the different pandemic waves. Results: Among 746 qualifying hospitals, admissions to hospital were significantly reduced after the covid-19 pandemic compared with before the pandemic for acute ischemic stroke (-13.59% (95% confidence interval-13.77% to -13.41%) and acute myocardial infarction (-17.20% (-17.39% to -17.01%)), as well as intravenous thrombolysis (-9.47% (-9.99% to -9.02%)), any percutaneous coronary intervention (-17.89% (-18.06% to -17.71%)), and percutaneous coronary intervention for acute myocardial infarction (-14.36% (-14.59% to -14.12%)). During the first year of the pandemic versus the previous year, the odds of mortality in hospital for acute ischemic stroke were 9.00% higher (3.51% v 3.16%; ratio of the means 1.09 (95% confidence interval (1.03 to 1.15); P=0.0013) and for acute myocardial infarction were 18.00% higher (4.81% v 4.29%; ratio of the means 1.18 (1.13 to 1.23); P<0.0001). Conclusions: We observed substantial decreases in admissions to hospital with acute ischemic stroke and acute myocardial infarction, but an increase in mortality in hospital throughout the first year of the pandemic. Public health interventions are needed to prevent these reductions in future pandemics.

Acute hepatitis of unknown origin in children: Behind the statistics.

Since April 2022, the world has been witnessing a rapidly spreading outbreak of acute hepatitis of unknown origin in children < 16 years old that has affected several countries around the world. Most of the cases have presented with the clinical picture of severe hepatitis that has led to resorting to liver transplantation in several cases. Despite the numerous theories that have been suggested on the possible underlying etiologies of the outbreak, an association with hepatitis A-E viruses and a link to COVID-19 vaccines have been excluded. Adenovirus serotype 41 has been detected in numerous cases, which makes it the most likely underlying cause of the disease. Nevertheless, other hypotheses are being investigated to justify the severity of the clinical picture, which is not typical of this type of virus. This review aims to summarize the current knowledge about the outbreak, highlight the suggested working hypotheses, and report the public health measures undertaken to tackle the outbreak.

Tourism Demand Prediction after COVID-19 with Deep Learning Hybrid CNN–LSTM—Case Study of Vietnam and Provinces

The tourism industry experienced a positive increase after COVID-19 and is the largest segment in the foreign exchange contribution in developing countries, especially in Vietnam, where China has begun reopening its borders and lifted the pandemic limitation on foreign travel. This research proposes a hybrid algorithm, combined convolution neural network (CNN) and long short-term memory (LSTM), to accurately predict the tourism demand in Vietnam and some provinces. The number of new COVID-19 cases worldwide and in Vietnam is considered a promising feature in predicting algorithms, which is novel in this research. The Pearson matrix, which evaluates the correlation between selected features and target variables, is computed to select the most appropriate input parameters. The architecture of the hybrid CNN–LSTM is optimized by utilizing hyperparameter fine-tuning, which improves the prediction accuracy and efficiency of the proposed algorithm. Moreover, the proposed CNN–LSTM outperformed other traditional approaches, including the backpropagation neural network (BPNN), CNN, recurrent neural network (RNN), gated recurrent unit (GRU), and LSTM algorithms, by deploying the K-fold cross-validation methodology. The developed algorithm could be utilized as the baseline strategy for resource planning, which could efficiently maximize and deeply utilize the available resource in Vietnam.

Depression, anxiety, and stress among frontline health workers during the second wave of COVID-19 in southern Vietnam: A cross-sectional survey

Health workers around the world have taken on massive frontline roles in the fight against COVID-19, often under intense pressure and in the face of uncertainty. In this study, we determined the rates of depression, anxiety, stress and related factors among health workers in COVID-19 designated hospitals in southern Vietnam during the second wave of COVID-19. From July-September 2020, we collected self-administered surveys from 499 health workers in 14 hospitals that were designated for the care and treatment of patients with COVID-19. The survey included sections on demographics, co-morbid health conditions, symptoms experienced during patient care, a depression, anxiety and stress assessment (DASS-21), and other related factors. We used logistic regression models to identify factors associated with depression, anxiety and stress, and adjusted for confounding factors. 18%, 11.5%, 7.7% of participants had symptoms of depression, anxiety, and stress, respectively with the majority at mild and moderate levels. The risk factors for increased mental health impact included long working hours, experiencing physical symptoms, fear of transmission to family, COVID-19 related stigma, and worry when watching media about COVID-19. Psychological counseling and training in infection prevention were protective factors that reduced the risk of mental health problems. Further exploration of the association between physical symptoms experienced by health workers and mental health may guide interventions to improve health outcomes. More routine COVID-19 testing among health workers could reduce anxieties about physical symptoms and alleviate the fear of transmitting COVID-19 to family and friends. Medical institutions need to ensure that health workers have access to basic trainings prior to initiation of work, and mental health support during the pandemic and into the future.

The global response: How cities and provinces around the globe tackled COVID-19 outbreaks in 2021—Authors' reply

Third, regarding the inclusion of RT-PCR and RATs in the laboratory tests, as mentioned in the text (page 9 of manuscript), "All cities/provinces except Tokyo (Table 5), relied on using mass testing strategies comprising both rapid antigen tests and standard RT-PCRs.” The number of laboratory tests was gathered from official websites of relevant national reporting authorities. As mentioned in the footnotes of the table, the WHO "had previously suggested a positivity rate of around 3-12% as a general benchmark of adequate testing, along with recommending that test positivity should remain at 5% or lower for 14 days before regions reopen." The point raised by the Ngo et al., has already been addressed in the manuscript.

Trial of Endovascular Therapy for Acute Ischemic Stroke with Large Infarct.

BACKGROUND: The role of endovascular therapy for acute stroke with a large infarction has not been extensively studied in differing populations. METHODS: We conducted a multicenter, prospective, open-label, randomized trial in China involving patients with acute large-vessel occlusion in the anterior circulation and an Alberta Stroke Program Early Computed Tomography Score of 3 to 5 (range, 0 to 10, with lower values indicating larger infarction) or an infarct-core volume of 70 to 100 ml. Patients were randomly assigned in a 1:1 ratio within 24 hours from the time they were last known to be well to undergo endovascular therapy and receive medical management or to receive medical management alone. The primary outcome was the score on the modified Rankin scale at 90 days (scores range from 0 to 6, with higher scores indicating greater disability), and the primary objective was to determine whether a shift in the distribution of the scores on the modified Rankin scale at 90 days had occurred between the two groups. Secondary outcomes included scores of 0 to 2 and 0 to 3 on the modified Rankin scale. The primary safety outcome was symptomatic intracranial hemorrhage within 48 hours after randomization. RESULTS: A total of 456 patients were enrolled; 231 were assigned to the endovascular-therapy group and 225 to the medical-management group. Approximately 28% of the patients in both groups received intravenous thrombolysis. The trial was stopped early owing to the efficacy of endovascular therapy after the second interim analysis. At 90 days, a shift in the distribution of scores on the modified Rankin scale toward better outcomes was observed in favor of endovascular therapy over medical management alone (generalized odds ratio, 1.37; 95% confidence interval, 1.11 to 1.69; P = 0.004). Symptomatic intracranial hemorrhage occurred in 14 of 230 patients (6.1%) in the endovascular-therapy group and in 6 of 225 patients (2.7%) in the medical-management group; any intracranial hemorrhage occurred in 113 (49.1%) and 39 (17.3%), respectively. Results for the secondary outcomes generally supported those of the primary analysis. CONCLUSIONS: In a trial conducted in China, patients with large cerebral infarctions had better outcomes with endovascular therapy administered within 24 hours than with medical management alone but had more intracranial hemorrhages. (Funded by Covidien Healthcare International Trading [Shanghai] and others; ANGEL-ASPECT ClinicalTrials.gov number, NCT04551664.).

One Health Surveillance Highlights Circulation of Viruses with Zoonotic Potential in Bats, Pigs, and Humans in Viet Nam.

A One Health cross-sectoral surveillance approach was implemented to screen biological samples from bats, pigs, and humans at high-risk interfaces for zoonotic viral spillover for five viral families with zoonotic potential in Viet Nam. Over 1600 animal and human samples from bat guano harvesting sites, natural bat roosts, and pig farming operations were tested for coronaviruses (CoVs), paramyxoviruses, influenza viruses, filoviruses and flaviviruses using consensus PCR assays. Human samples were also tested using immunoassays to detect antibodies against eight virus groups. Significant viral diversity, including CoVs closely related to ancestors of pig pathogens, was detected in bats roosting at the human-animal interfaces, illustrating the high risk for CoV spillover from bats to pigs in Viet Nam, where pig density is very high. Season and reproductive period were significantly associated with the detection of bat CoVs, with site-specific effects. Phylogeographic analysis indicated localized viral transmission among pig farms. Our limited human sampling did not detect any known zoonotic bat viruses in human communities living close to the bat cave and harvesting bat guano, but our serological assays showed possible previous exposure to Marburg virus-like (Filoviridae), Crimean-Congo hemorrhagic fever virus-like (Bunyaviridae) viruses and flaviviruses. Targeted and coordinated One Health surveillance helped uncover this viral pathogen emergence hotspot.

Willingness and attitudes of parents towards COVID-19 vaccines for children in Vietnam.

BACKGROUND: With an increasing need for COVID-19 vaccination around the globe, we aim to investigate willingness and attitudes of parents regarding COVID-19 vaccines for children in Vietnam. METHOD: A 24-item online survey was conducted among 602 parents and legal guardians of children under 18 years of age. RESULTS: There were 82.6% of parents willing to vaccinate their children. The principal reasons for willingness were: the need for vaccination to control the COVID-19 pandemic and to reduce the risk of SARS-CoV-2 transmission when children return to school. The most common reason for refusing COVID-19 vaccine was the concern about vaccine side effects in children (73.3%), followed by perceived lack of scientific research on COVID-19 vaccines in children (31.4%). The main associated factors with increased willingness of parents towards childhood COVID-19 vaccine were: willingness to allow children to participate in a clinical vaccine trial (aOR = 3.58); possible increase in COVID-19-related mortality (aOR = 3.69); and positive media information regarding COVID-19 vaccine (aOR = 2.04). Noteworthy, higher educational status of parents was associated with decreased willingness for childhood COVID-19 vaccine (aOR = 0.26). CONCLUSION: Understanding reasons of parents to accept childhood COVID-19 vaccines will help future plans in COVID-19 vaccination program. A larger sample size study at a national level is needed to verify the results.

A Single-Administration Microneedle Skin Patch for Multi-Burst Release of Vaccine against SARS-CoV-2.

The necessity for multiple injections and cold-chain storage has contributed to suboptimal vaccine utilization, especially in pandemic situations. Thermally-stable and single-administration vaccines hold a great potential to revolutionize the global immunization process. Here, a new approach to thermally stabilize protein-based antigens is presented and a new high-throughput antigen-loading process is devised to create a single-administration, pulsatile-release microneedle (MN) patch which can deliver a recombinant SARS-CoV-2 S1-RBD protein-a model for the COVID-19 vaccine. Nearly 100% of the protein antigen could be stabilized at temperatures up to 100 °C for at least 1 h and at an average human body temperature (37 °C) for up to 4 months. Arrays of the stabilized S1-RBD formulations can be loaded into the MN shells via a single-alignment assembly step. The fabricated MNs are administered at a single time into the skin of rats and induce antibody response which could neutralize authentic SARS-CoV-2 viruses, providing similar immunogenic effect to that induced by multiple bolus injections of the same antigen stored in conventional cold-chain conditions. The MN system presented herein could offer the key solution to global immunization campaigns by avoiding low patient compliance, the requirement for cold-chain storage, and the need for multiple booster injections.

Spatiotemporal Evolution of SARS-CoV-2 Alpha and Delta Variants during Large Nationwide Outbreak of COVID-19, Vietnam, 2021.

We analyzed 1,303 SARS-CoV-2 whole-genome sequences from Vietnam, and found the Alpha and Delta variants were responsible for a large nationwide outbreak of COVID-19 in 2021. The Delta variant was confined to the AY.57 lineage and caused >1.7 million infections and >32,000 deaths. Viral transmission was strongly affected by nonpharmaceutical interventions.

Assessing cognitive decline in Vietnamese older adults using the Montreal Cognitive Assessment-Basic (MoCA-B) and Informant Questionnaire on Cognitive Decline in the Elderly (IQCODE) during the COVID-19 pandemic: A feasibility study.

Objectives: The lack of cognitive assessment tools suitable for people with minimal formal education is a barrier to identify cognitive impairment in Vietnam. Our aims were to (i) evaluate the feasibility of conducting the Montreal Cognitive Assessment-Basic (MoCA-B) and Informant Questionnaire On Cognitive Decline in the Elderly (IQCODE) remotely on the Vietnamese older adults, (ii) examine the association between the two tests, (iii) identify demographic factors correlated with these tools. Methods: The MoCA-B was adapted from the original English version, and a remote testing procedure was conducted. One hundred seventy-three participants aged 60 and above living in the Vietnamese southern provinces were recruited via an online platform during the COVID-19 pandemic. Results: IQCODE results showed that the proportions of rural participants classified as having mild cognitive impairment and dementia were substantially higher than those in urban areas. Levels of education and living areas were associated with IQCODE scores. Education attainment was also the main predictor of MoCA-B scores (30% of variance explained), with an average of 10.5 points difference between those with no formal education and those who attended university. Conclusions: It is feasible to administer the IQCODE and MoCA-B remotely in the Vietnamese older population. Education attainment played a stronger role in predicting MoCA-B scores than IQCODE, suggesting the influence of this factor on MoCA-B scores. Further study is needed to develop socio-culturally appropriate cognitive screening tests for the Vietnamese population.

Predicting the severity of COVID-19 patients using the CD24-CSF1R index in whole blood samples.

Coronavirus disease 2019 (COVID-19), caused by SARS-CoV-2, has become one of the most serious public health crises worldwide. Most infected people are asymptomatic but are still able to spread the virus. People with mild or moderate illnesses are likely to recover without hospitalization, while critically ill patients face a higher risk of organ injury or even death. In this study, we aimed to identify a novel biomarker that can predict the severity of COVID-19 patients. Clinical information and RNA-seq data of leukocytes from whole blood samples with and without a COVID-19 diagnosis (n = 100 and 26, respectively) were retrieved from the National Center for Biotechnology Information Gene Expression Omnibus database. Raw data were processed using the Transcripts Per Million (TPM) method and then transformed using log2 (TPM+1) for normalization. The CD24-CSF1R index was established. Violin plots, Kaplan-Meier curves, ROC curves, and multivariate Cox proportional hazards regression analyses were performed to evaluate the prognostic value of the established index. The CD24-CSF1R index was significantly associated with ICU admission (n = 50 ICU, 50 non-ICU) and ventilatory status (n = 42 ventilation, 58 non-ventilation) with p = 4.186e-11 and p = 1.278e-07, respectively. The ROC curve produced a relatively accurate prediction of ICU admission with an AUC of 0.8524. Additionally, patients with a high index had significantly fewer mechanical ventilation-free days than patients with a low index (p = 6.07e-07). Furthermore, the established index showed a strong prognostic ability for the risk of using a ventilator in the multivariate Cox regression model (p < 0.001). The CD24-CSF1R index was significantly associated with COVID-19 severity. The established index could have potential implications for prognosis, disease severity stratification, and clinical management.

Side effects following first dose of COVID-19 vaccination in Ho Chi Minh City, Vietnam.

Vaccines are strongly recommended globally as an effective measure to prevent serious illness from and spread of COVID-19. Concerns about safety following vaccination continue to be the most common reason that people do not accept the vaccine. This retrospective study was carried out on 4341 people who received the first dose of ChAdOx1 nCoV-19, BBIBP-CorV, or mRNA-1273 vaccine at Jio Health Clinic in Ho Chi Minh City, Vietnam. Post-injection side effects were either reported by participants or actively collected by health care staff by means of telemedicine. Local side effects were reported by 35.5% of all individuals, with pain being the most common symptom (33.3%). Systemic side effects were reported by 44.2% of individuals, with fever (25.3%) and fatigue (21.4%) being the most common. Age ≤60 years, female gender, and ChAdOx1 nCoV-19 were significant independent risk factors for both local and systemic side effects, while a history of allergy was significant as a risk factor for local side effects. A total of 43 individuals (1.0%) reported concerning symptoms of rare severe complications, which were addressed and treated by physicians via Jio Health app.

Mapping patient pathways and understanding clinical decision-making in dengue management to inform the development of digital health tools.

BACKGROUND: Dengue is a common viral illness and severe disease results in life-threatening complications. Healthcare services in low- and middle-income countries treat the majority of dengue cases worldwide. However, the clinical decision-making processes which result in effective treatment are poorly characterised within this setting. In order to improve clinical care through interventions relating to digital clinical decision-support systems (CDSS), we set out to establish a framework for clinical decision-making in dengue management to inform implementation. METHODS: We utilised process mapping and task analysis methods to characterise existing dengue management at the Hospital for Tropical Diseases, Ho Chi Minh City, Vietnam. This is a tertiary referral hospital which manages approximately 30,000 patients with dengue each year, accepting referrals from Ho Chi Minh city and the surrounding catchment area. Initial findings were expanded through semi-structured interviews with clinicians in order to understand clinical reasoning and cognitive factors in detail. A grounded theory was used for coding and emergent themes were developed through iterative discussions with clinician-researchers. RESULTS: Key clinical decision-making points were identified: (i) at the initial patient evaluation for dengue diagnosis to decide on hospital admission and the provision of fluid/blood product therapy, (ii) in those patients who develop severe disease or other complications, (iii) at the point of recurrent shock in balancing the need for fluid therapy with complications of volume overload. From interviews the following themes were identified: prioritising clinical diagnosis and evaluation over existing diagnostics, the role of dengue guidelines published by the Ministry of Health, the impact of seasonality and caseload on decision-making strategies, and the potential role of digital decision-support and disease scoring tools. CONCLUSIONS: The study highlights the contemporary priorities in delivering clinical care to patients with dengue in an endemic setting. Key decision-making processes and the sources of information that were of the greatest utility were identified. These findings serve as a foundation for future clinical interventions and improvements in healthcare. Understanding the decision-making process in greater detail also allows for development and implementation of CDSS which are suited to the local context.

SARS-CoV-2 Omicron and its current known unknowns: A narrative review.

The emergence of the SARS-CoV-2 Omicron variant (B.1.1.529) has created great global distress. This variant of concern shows multiple sublineages, importantly B.1.1.529.1 (BA.1), BA.1 + R346K (BA.1.1), and B.1.1.529.2 (BA.2), each with unique properties. However, little is known about this new variant, specifically its sub-variants. A narrative review was conducted to summarise the latest findings on transmissibility, clinical manifestations, diagnosis, and efficacy of current vaccines and treatments. Omicron has shown two times higher transmission rates than Delta and above ten times more infectious than other variants over a similar period. With more than 30 mutations in the spike protein's receptor-binding domain, there is reduced detection by conventional RT-PCR and rapid antigen tests. Moreover, the two-dose vaccine effectiveness against Delta and Omicron variants was found to be approximately 21%, suggesting an urgent need for a booster dose to prevent the possibility of breakthrough infections. However, the current vaccines remain highly efficacious against severe disease, hospitalisation, and mortality. Japanese preliminary lab data elucidated that the Omicron sublineage BA.2 shows a higher illness severity than BA.1. To date, the clinical management of Omicron remains unchanged, except for monoclonal antibodies. Thus far, only Bebtelovimab could sufficiently treat all three sub-variants of Omicron. Further studies are warranted to understand the complexity of Omicron and its sub-variants. Such research is necessary to improve the management and prevention of Omicron infection.