Download This Paper Open PDF in Browser Add Paper to My Library Share: Permalink Using these links will ensure access to this page indefinitely Copy URL Preprints with The Lancet is part of SSRN´s First Look, a place where journals identify content of interest prior to publication. Authors have opted in at submission to The Lancet family of journals to post their preprints on Preprints with The Lancet. The usual SSRN checks and a Lancet-specific check for appropriateness and transparency have been applied. Preprints available here are not Lancet publications or necessarily under review with a Lancet journal. These preprints are early stage research papers that have not been peer-reviewed. The findings should not be used for clinical or public health decision making and should not be presented to a lay audience without highlighting that they are preliminary and have not been peer-reviewed. For more information on this collaboration, see the comments published in The Lancet about the trial period, and our decision to make this a permanent offering, or visit The Lancet´s FAQ page, and for any feedback please contact preprints@lancet.com. Using Machine Learning to Identify Severe Dengue Cases in the Elderly and Non-Elderly in Southern Taiwan: A Step Toward International Collaboration on Dengue Syndromic Surveillance 47 Pages Posted: 27 Feb 2024 See all articles by Tzong-Shiann HoTzong-Shiann HoNational Cheng Kung University - National Cheng Kung University HospitalWei ChenNational Cheng Kung UniversityChin-Rur YangNational Taiwan UniversityChi-Hua YuNational Cheng Kung UniversityYu-Pei ChenMinistry of Health and Welfare (Taiwan)Chih-Huan ChungKuo General HospitalYa-Yun ChengNational Sun Yat-sen UniversityKo ChangKaohsiung Medical UniversityWei-Ting HsuNational Taiwan UniversitySingLek LiewNational Taiwan UniversityYin-Chuan ChaoNational Cheng Kung UniversityWen-Xuan HuangNational Cheng Kung UniversityChun-Ying WuNational Taiwan UniversityThomas Claudio Guillermo TsaiNational Taiwan UniversityMarie Yung-Chen WuNational Taiwan UniversityYa-Fan TsaoNational Taiwan UniversityChia-Chun ChangNational Taiwan UniversityIng LeeKaohsiung Chang Gung Memorial HospitalChwan-Chuen KingNational Taiwan University More... Abstract Background: High severity and case fatality rates among elderly dengue patients have become growing threats globally. Early recognition of severe dengue is crucial to saving lives.Methods: To evaluate the predictive factors leading to severe dengue in elderly patients, we enrolled 1,728 laboratory-confirmed dengue cases [including 267 severe (15.5%) and 1461 mild cases] from southern Taiwan: (1) two community hospitals and one tertiary medical center in 2015 outbreak in Tainan and (2) another medical center during outbreaks from 2002-2013 in Kaohsiung. Among these, 606 (35.1%) patients were elderly (≥ 65 years). The individual’s clinical symptoms/signs and laboratory findings of dengue were retrospectively collected from chart reviews. Data analyses were run using both epidemiological and machine learning approaches. We further applied the Synthetic Minority Oversampling Technique (SMOTE) to solve the imbalance between severe and non-severe cases. Then, we used a random forest model to select essential features to differentiate between the clinical manifestations of severe and mild cases.Findings: Our model using triage symptoms in the elderly and non-elderly achieved 97% accuracy in training and 95% accuracy in testing on predicting severe dengue. Hemorrhage ranked the highest symptom in predicting severe dengue. Fever was important for diagnosing severe dengue in the elderly, whereas GI-related symptoms were significant for predicting severe dengue in the non-elderly. Lastly, we developed a cloud service to help clinicians detect severe dengue cases earlier.Interpretation: The machine learning model involving 15 triage symptoms, 43 accessory features, and four comorbidities predicted severe dengue cases in the elderly with > 95% accuracy.Funding: The study was sponsored by the National Health Research Institute (NHRI), National Science and Technology Council [NSTC, formerly the Ministry of Science and Technology (MOST) in Taiwan], and National Cheng Kung University Hospital (NCKUH) (NHRI Grant numbers: NHRI 110A1-MRCO-01212103, NHRI-11A1-MRCO-01222202, NHRI-12A1-MRCO-0123-2301, NSTC/MOST Grant Numbers: MOST 105-2815-C-002-059-B, NSTC-111-2327-B-006, NSTC 111-2327-B-005; NCKUH Grant number: NCKUH-11002018) in Taiwan.Declaration of Interest: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.Ethical Approval: Study protocols were reviewed, approved, and executed following the Institutional Review Boards of the National Health Research Institute (NHRI), Miaoli County, Taiwan (NHRI: EC1051108-R2), National Taiwan University (NTU:202212HM042, 202103HM008, 202112HM050), National Cheng Kung University Hospital (NCKUH: A-BR-101-140), Tainan Hospital, Ministry of Health and Welfare (MHW-TNH: 17-017), and Kaohsiung Chang Gung Memorial Hospital (KHCGMH: 202200783B0) Keywords: severe dengue, syndromic surveillance, elderly, comorbidity, machine learning, Taiwan Suggested Citation: Suggested Citation Ho, Tzong-Shiann and Chen, Wei and Yang, Chin-Rur and Yu, Chi-Hua and Chen, Yu-Pei and Chung, Chih-Huan and Cheng, Ya-Yun and Chang, Ko and Hsu, Wei-Ting and Liew, SingLek and Chao, Yin-Chuan and Huang, Wen-Xuan and Wu, Chun-Ying and Tsai, Thomas Claudio Guillermo and Wu, Marie Yung-Chen and Tsao, Ya-Fan and Chang, Chia-Chun and Lee, Ing and King, Chwan-Chuen, Using Machine Learning to Identify Severe Dengue Cases in the Elderly and Non-Elderly in Southern Taiwan: A Step Toward International Collaboration on Dengue Syndromic Surveillance. Available at SSRN: https://ssrn.com/abstract=4736609 Tzong-Shiann Ho National Cheng Kung University - National Cheng Kung University Hospital ( email ) TainanTaiwan Wei Chen National Cheng Kung University ( email ) No.1, University RoadTainanTaiwan Chin-Rur Yang National Taiwan University ( email ) Chi-Hua Yu (Contact Author) National Cheng Kung University ( email ) Yu-Pei Chen Ministry of Health and Welfare (Taiwan) ( email ) Chih-Huan Chung Kuo General Hospital ( email ) Ya-Yun Cheng National Sun Yat-sen University ( email ) 70 Lien-hai Rd.Kaohsiung, 80743Taiwan Ko Chang Kaohsiung Medical University ( email ) Wei-Ting Hsu National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan SingLek Liew National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan Yin-Chuan Chao National Cheng Kung University ( email ) No.1, University RoadTainanTaiwan Wen-Xuan Huang National Cheng Kung University ( email ) No.1, University RoadTainanTaiwan Chun-Ying Wu National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan Thomas Claudio Guillermo Tsai National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan Marie Yung-Chen Wu National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan Ya-Fan Tsao National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan Chia-Chun Chang National Taiwan University ( email ) 1 Sec. 4, Roosevelt RoadTaipei 106, 106Taiwan Ing Lee Kaohsiung Chang Gung Memorial Hospital ( email ) Chwan-Chuen King National Taiwan University ( email ) Download This Paper Open PDF in Browser Please enable JavaScript to view the comments powered by Disqus. Click here to go to TheLancet.com Go to TheLancet.com Paper statistics Downloads 6 Abstract Views 23 75 References PlumX Metrics Preprints with The Lancet Specialties All Journals (16541) Allergy & Immunology (1064) Anaesthesia & Analgesia (51) Cardiology & Vascular Medicine (1159) Child & Adoloscence Health (1297) Critical Care (577) Dermatology (104) Digital Health (805) Endocrinology (952) Gastroenterology (958) Genetics & Genomics (2200) Geriatrics (201) Global Health (1281) Haematology (652) Infectious Diseases (5577) Nephrology (340) Neurology (1463) Nutrition (496) Obstetrics & Gynaecology (767) Oncology (3344) Ophthalmology (228) Otolaryngology (112) Planetary Health (256) Primary Care (853) Psychiatry (1080) Public Health (3541) Radiology & Medical Imaging (451) Respiratory Medicine (696) Rheumatology & Orthopaedics (603) Surgery (398) Urology (97) Related eJournals Artificial Intelligence eJournal Follow Artificial Intelligence eJournal Subscribe to this fee journal for more curated articles on this topic FOLLOWERS 392 PAPERS 10,391 Preprints with The Lancet Follow Preprints with The Lancet Subscribe to this free journal for more curated articles on this topic FOLLOWERS 256 PAPERS 16,541 Global Health eJournal Follow Global Health eJournal Subscribe to this fee journal for more curated articles on this topic FOLLOWERS 58 PAPERS 1,321 Feedback Feedback to SSRN Feedback (required) Email (required) Submit If you need immediate assistance, call 877-SSRNHelp (877 777 6435) in the United States, or +1 212 448 2500 outside of the United States, 8:30AM to 6:00PM U.S. Eastern, Monday - Friday.

Taiwan's dengue cases vary annually, peaking in infrequent epidemics, which differ substantially from the Global Burden of Disease Study's (GBD's) projections. Although the GBD study provides invaluable insights into global health trends, its modelling approach fails to capture the dynamic change of dengue transmission.

Taiwan’s experience with severe acute respiratory syndrome coronavirus (SARS-CoV) in 2003 guided its development of strategies to defend against SARS-CoV-2 in 2020, which enabled the successful control of Coronavirus disease 2019 (COVID-19) cases from 2020 through March 2021. However, in late-April 2021, the imported Alpha variant began to cause COVID-19 outbreaks at an exceptional rate in Taiwan. In this study, we aimed to determine what epidemiological conditions enabled the SARS-CoV-2 Alpha variant strains to become dominant and decline later during a surge in the outbreak. In conjunction with contact-tracing investigations, we used our bioinformatics software, CoVConvert and IniCoV, to analyze whole-genome sequences of 101 Taiwan Alpha strains. Univariate and multivariable regression analyses revealed the epidemiological factors associated with viral dominance. Univariate analysis showed the dominant Alpha strains were preferentially selected in the surge’s epicenter (p = 0.0024) through intensive human-to-human contact and maintained their dominance for 1.5 months until the Zero-COVID Policy was implemented. Multivariable regression found that the epidemic periods (p = 0.007) and epicenter (p = 0.001) were two significant factors associated with the dominant virus strains spread in the community. These dominant virus strains emerged at the outbreak’s epicenter with frequent human-to-human contact and low vaccination coverage. The Level 3 Restrictions and Zero-COVID policy successfully controlled the outbreak in the community without city lockdowns. Our integrated method can identify the epidemiological conditions for emerging dominant virus with increasing epidemiological potential and support decision makers in rapidly containing outbreaks using public health measures that target fast-spreading virus strains.

Intro: Severe outbreaks in Taiwan during 1987-2015 had always started from failure in detecting early dengue cases. However, the extreme low numbers of indigenous dengue cases in southern Taiwan during 2016-2022 have been succeeded with early efforts through integrated surveillance systems, including, free charge NS1 screening of virological surveillance for patients with dengue-like illness, fever surveillance at the airport and entomological surveillance looking for mosquito breeding sites. Recently, we used machine learning models to assist in finding lab.-confirmed dengue cases and predicting severe dengue cases in the elderly. Methods: We investigated factors that were associated with severe dengue outbreaks in southern Taiwan and used geographical information system (GIS) to plot spatial distribution of laboratory-confirmed dengue cases. Findings: Six epidemiological findings can direct prevention and control policies. (1) The onset of epidemics begins with imported dengue cases and appropriate weather conditions (high temperature, low humidity) enlightening to educate the symptoms/signs of dengue for travelers is important. (2) DENV-2, had resulted in most severe epidemics. (3) Once DENV- 2/DENV-3 circulates too long or severe cases occur, fatal dengue cases will increase alongside the prolongation of epidemic wave. (4) In areas with high population densities, either longer epidemic wave or higher transmission intensity can result in greater percentage of severe dengue cases. However, in areas with low population densities, both conditions can lead to greater percentage of severe dengue cases Therefore., interrupting DENV transmission efficiently can reduce severe/fatal cases. (5) Severe dengue cases frequently occur in tempo-spatial dengue clusters implying the importance to find out the sources of infection. The overlapping areas between high mosquito indices and dengue clusters facilitate transmission persistence. (6) Source reduction is more effective than insecticide-spraying. Conclusion: integrated surveillance systems, immediate interrupting transmission, and identifying mosquito breeding sites all together can reduce epidemic severity and thus promoting global health.

Background: The objective of this study was to evaluate the effects of prior-infection and repeated vaccination on post-vaccination antibody titers. Methods: A(H1N1)pdm09 strain was included in 2009 pandemic monovalent, 2010-2011, and 2011- 2012 trivalent influenza vaccines (MIVpdm09, TIV10/11, TIV11/12) in Taiwan. During the 2011-2012 influenza season, we conducted a prospective sero-epidemiological cohort study among schoolchildren from grades 1 - 6 in the two elementary schools in Taipei with documented A(H1N1)pdm09 vaccination records since 2009. Serum samples were collected at pre-vaccination, 1-month, and 4-months postvaccination (T1, T2, T3). Anti-A(H1N1)pdm09 hemagglutination inhibition titers (HI-Ab-titers) were examined. We also investigated the impact of four vaccination histories [(1) no previous vaccination (None), (2) vaccinated in 2009-2010 season (09v), (3) vaccinated in 2010-2011 season (10v), and (4) vaccinated consecutively in 2009-2010 and 2010-2011 seasons (09v + 10v)] and pre-vaccination HI-Ab levels on post-vaccination HI-Ab responses as well as adjusted vaccine effectiveness (aVE) against serologically-defined infection from T2 to T3. Results: TIV11/12 had zero serious adverse events reported. A(H1N1)pdm09 strain in TIV11/12 elicited seroprotective Ab-titers in 98% of children and showed promising protection (aVE: 70.3% [95% confidence interval (CI): 51.0-82.1%]). Previously unvaccinated but infected children had a 3.96 times higher T2 geometric mean titer (T2-GMT) of HI-Ab than those naive to A(H1N1)pdm09 (GMT [95% CI]: 1039.7[585.3- 1845.9] vs. 262.5[65.9-1045], p = 0.046). Previously vaccinated children with seroprotective T1-Ab-titers had a higher T2-GMT and a greater aVE than those with non-seroprotective T1-Ab-titers. Repeatedly vaccinated children had lower T2-GMT than those receiving primary doses of TIV11/12. However, after controlling prior infection and T1-Ab-titers, differences in T2-GMT among the four vaccination histories became insignificant (p = 0.16). Conclusion: This study supports the implementation of annual mass-vaccination with A(H1N1)pdm09 in schoolchildren for three consecutive influenza seasons when vaccine and circulating strains were well matched, and found that prior infection and pre-vaccination HI-Ab levels positively impacted postvaccination HI-Ab responses.

OBJECTIVES:To evaluate class suspension and mass vaccination implemented among Taipei schoolchildren during the 2009 influenza pandemic and investigate factors affecting antibody responses. METHODS:We conducted 2 cohort studies on: (1) 972 schoolchildren from November 2009-March 2010 to evaluate pandemic policies and (2) 935 schoolchildren from November 2011-March 2012 to verify factors in antibody waning. Anti-influenza H1N1pdm09 hemagglutination inhibition antibodies (HI-Ab) were measured from serum samples collected before vaccination, and at 1 and 4 months after vaccination. Factors affecting HI-Ab responses were investigated through logistic regression and generalized estimating equation. RESULTS:Seroprevalence of H1N1pdm09 before vaccination was significantly higher among schoolchildren who experienced class suspensions than those who did not (59.6% vs 47.5%, p<0.05). Participating in after-school activities (adjusted odds ratio [aOR]=2.47, p=0.047) and having ≥3 hours per week of exercise (aOR=2.86, p=0.019) were significantly correlated with H1N1pdm09 infection. Two doses of the H1N1pdm09 vaccine demonstrated significantly better antibody persistence than 1 dose (HI-Ab geometric mean titer: 132.5 vs 88.6, p=0.047). Vaccine effectiveness after controlling for preexisting immunity was 86% (32%-97%). Exercise ≥3 hours per week and preexisting immunity were significantly associated with antibody waning/maintenance. CONCLUSIONS:This study is the first to show that exercise and preexisting immunity may affect antibody waning. Further investigation is needed to identify immune correlates of protection.

Emerging infectious diseases (EIDs), including the latest COVID-19 pandemic, have emerged and raised global public health crises in recent decades. Without existing protective immunity, an EID may spread rapidly and cause mass casualties in a very short time. Therefore, it is imperative to identify cases with risk of disease progression for the optimized allocation of medical resources in case medical facilities are overwhelmed with a flood of patients. This study has aimed to cope with this challenge from the aspect of preventive medicine by exploiting machine learning technologies. The study has been based on 83,227 hospital admissions with influenza-like illness and we analysed the risk effects of 19 comorbidities along with age and gender for severe illness or mortality risk. The experimental results revealed that the decision rules derived from the machine learning based prediction models can provide valuable guidelines for the healthcare policy makers to develop an effective vaccination strategy. Furthermore, in case the healthcare facilities are overwhelmed by patients with EID, which frequently occurred in the recent COVID-19 pandemic, the frontline physicians can incorporate the proposed prediction models to triage patients suffering minor symptoms without laboratory tests, which may become scarce during an EID disaster. In conclusion, our study has demonstrated an effective approach to exploit machine learning technologies to cope with the challenges faced during the outbreak of an EID.

Background The Delta and Omicron variants of SARS-CoV-2 are currently responsible for breakthrough infections due to waning immunity. We report phase I/II trial results of UB-612, a multitope subunit vaccine containing S1-RBD-sFc protein and rationally designed promiscuous peptides representing sarbecovirus conserved helper T cell and cytotoxic T lymphocyte epitopes on the nucleocapsid (N), membrane (M), and spike (S2) proteins. Method We conducted a phase I primary 2-dose (28 days apart) trial of 10, 30, or 100 μg UB-612 in 60 healthy young adults 20 to 55 years old, and 50 of them were boosted with 100 μg of UB-612 approximately 7 to 9 months after the second dose. A separate placebo-controlled and randomized phase II study was conducted with 2 doses of 100 μg of UB-612 (n = 3,875, 18–85 years old). We evaluated interim safety and immunogenicity of phase I until 14 days after the third (booster) dose and of phase II until 28 days after the second dose. Results No vaccine-related serious adverse events were recorded. The most common solicited adverse events were injection site pain and fatigue, mostly mild and transient. In both trials, UB-612 elicited respective neutralizing antibody titers similar to a panel of human convalescent sera. The most striking findings were long-lasting virus-neutralizing antibodies and broad T cell immunity against SARS-CoV-2 variants of concern (VoCs), including Delta and Omicron, and a strong booster-recalled memory immunity with high cross-reactive neutralizing titers against the Delta and Omicron VoCs. Conclusion UB-612 has presented a favorable safety profile, potent booster effect against VoCs, and long-lasting B and broad T cell immunity that warrants further development for both primary immunization and heterologous boosting of other COVID-19 vaccines. Trial Registration ClinicalTrials.gov: NCT04545749, NCT04773067, and NCT04967742. Funding UBI Asia, Vaxxinity Inc., and Taiwan Centers for Disease Control, Ministry of Health and Welfare.

Outbreaks of avian influenza virus (AIV) have raised public concerns recently. Airborne AIV has been evaluated in live poultry markets and case farms; however, no study has discussed airborne AIV in ambient air in the winter habitats of migratory birds. Therefore, this study aimed to evaluate airborne AIV, specifically H5, H7, and H9, in a critical winter habitat of migratory birds and assess the factors influencing airborne AIV transmission in ambient air to provide novel insights into the epidemiology of avian influenza. A total of 357 ambient air samples were collected in the Aogu Wetland, Taiwan, Republic of China, between October 2017 and December 2019 and analyzed using quantitative real-time polymerase chain reaction. The effects of environmental factors including air pollutants, meteorological factors, and the species of the observed migratory birds on the concentration of airborne AIV were also analyzed. To our knowledge, this is the first study to investigate the relationship between airborne AIV in ambient air and the influence factors in the winter habitats of migratory birds, demonstrating the benefits of environmental sampling for infectious disease epidemiology. The positive rate of airborne H7 (12%) was higher than that of H5 (8%) and H9 (10%). The daily mean temperature and daily maximum temperature had a significant negative correlation with influenza A, H7, and H9. Cold air masses and bird migration were significantly associated with airborne H9 and H7, respectively. In addition, we observed a significant correlation between AIV and the number of pintails, common teals, Indian spot-billed ducks, northern shovelers, Eurasian wigeons, tufted ducks, pied avocets, black-faced spoonbills, and great cormorants. In conclusion, we demonstrated the potential for alternative surveillance approaches (monitoring bird species) as an indicator for influenza-related risks and identified cold air masses and the presence of specific bird species as potential drivers of the presence and/or the airborne concentration of AIV.

Emerging infectious diseases (EIDs), including the latest COVID-19 pandemic, have emerged and raised global public health crises in recent decades. Without existing protective immunity, an EID may spread rapidly and cause mass casualties in a very short time. Therefore, it is imperative to identify cases with risk of disease progression for the best allocation of medical resources in case medical facilities are overwhelmed with a flood of patients. This study aimed to exploit machine learning technologies to cope with this challenge. The study was based on 83,227 hospital admissions with influenza-like illness and we analysed the risk effects of 19 comorbidities along with age and gender for severe illness or mortality risk. The experimental results revealed that the conventional decision tree (DT) models built with only 6 features, including age, gender, and four comorbidities, delivered the same level of prediction accuracy as the state-of-the-art deep neural network models built with 18 features. Accordingly, we further studied how to exploit the DT models with different sensitivity levels to determine patient triage and optimize medical resource allocation in different stages of an EID disaster to aid the frontline clinicians and policy-makers. In conclusion, our study demonstrated an approach to exploit machine learning technologies to cope with the challenges during the outbreak of an EID.

During 2012-2017, a total of 1,144 highly pathogenic avian influenza (HPAI) H5 outbreaks were reported in Taiwan. We conjectured the current 3-km radius of the post-outbreak containment policy could fail to effectively alleviate the current ongoing epidemics of HPAI H5 in Taiwan. The high intensity of localized transmission of HPAI H5 at certain focal hotspots was identified to follow the spatial distribution of poultry-raising locations through our hotspot analyses on the HPAI H5 outbreak locations from 2015 to 2017. We then applied 3-, 5- and 7-km circular buffer zones to 15,444 registered poultry-raising locations to inspect the characteristics of the poultry-raising neighbourhood. Three spatial regression models using Bayesian inference were established to infer the risks attributable to poultry-raising characteristics in the corresponding buffer areas. The different buffer radii were treated as a sensitivity analysis of the influential range of neighbouring farms on the HPAI H5 outbreak occurrence, so as to evaluate the effective radius for post-outbreak containment. Evidence showed that the risks of outbreak occurrence were associated with increasing numbers of poultry-raising locations in both 3-km (relative risk [RR] 1.005, 95% confidence interval [CI] 1.002-1.008) and 5-km buffer areas (RR 1.005, 95% CI 1.004-1.007), whereas in the 7-km buffer model, no association between densely populated locations and increasing risks of outbreaks was observed (RR 1.000, 95% CI 0.999-1.001). Therefore, an extension to a 7-km radius for the post-outbreak containment policy (rather than a 3-km radius as in the current policy) is recommended to effectively mitigate further spreading of HPAI H5 outbreaks among neighbouring farms. Overall, we demonstrated that the densely populated locations with multiple poultry species raised in proximity as defined with 3-, 5- and 7-km buffer areas facilitated H5 HPAI outbreak diffusion and shaped the scale of HPAI H5 epidemics in Taiwan.

Objectives: To evaluate the prevalence of infection prevention behaviors in Taiwan—wearing facemasks and alcohol-based hand hygiene (AHH)—and compare their practice rates during SARS and COVID-19. Methods: We surveyed 2328 Taiwanese from July 29 to August 6, 2020, assessing demographics, information sources, and preventive behaviors during the 2003 SARS outbreaks, 2009 pandemic influenza H1N1, COVID-19, and with post-survey intentions. Characteristics associated with the practice of preventive behaviors in 2020 were identified through logistic regression. Results: Preventive behaviors were conscientiously practiced by 70.2% of participants. Compared with 2003 SARS/2009 H1N1, the percentages of facemask use (66.6% vs 99.2% [indoors], P < 0.001) and on-person AHH (44.2% vs 65.4% [hand sanitizers], P < 0.001) significantly increasedduring 2020 COVID-19. Highest adherence to preventive behaviors in 2020 was among females (adjusted odds ratio [aOR], 1.72), those receiving government COVID-19 information (aOR, 1.52), participants recruited from primary-care clinics (aOR, 1.43), and those who practiced AHH during 2003 SARS/2009 H1N1 (aOR, 1.37). Conclusions: Government leadership, healthcare providers risk communication, and public cooperation rapidly mitigated the spread of COVID-19 in Taiwan even before vaccination. Future global efforts must implement such population-based preventive behaviors at a level above the viral-transmission-threshold, particularly in areas with fast-spreading SARS-CoV-2 variants.

A shift in dengue cases toward the adult population, accompanied by an increased risk of severe cases of dengue in the elderly, has created an important emerging issue in the past decade. To understand the level of past DENV infection among older adults after a large dengue outbreak occurred in southern Taiwan in 2015, we screened 1498 and 2603 serum samples from healthy residents aged ≥ 40 years in Kaohsiung City and Tainan City, respectively, to assess the seroprevalence of anti-DENV IgG in 2016. Seropositive samples were verified to exclude cross-reaction from Japanese encephalitis virus (JEV), using DENV/JEV-NS1 indirect IgG ELISA. We further identified viral serotypes and secondary DENV infections among positive samples in the two cities. The overall age-standardized seroprevalence of DENV-IgG among participants was 25.77% in Kaohsiung and 11.40% in Tainan, and the seroprevalence was significantly higher in older age groups of both cities. Although the percentages of secondary DENV infection in Kaohsiung and Tainan were very similar (43.09% and 44.76%, respectively), DENV-1 and DENV-2 spanned a wider age range in Kaohsiung, whereas DENV-2 was dominant in Tainan. As very few studies have obtained the serostatus of DENV infection in older adults and the elderly, this study highlights the need for further investigation into antibody status, as well as the safety and efficacy of dengue vaccination in these older populations.

Background: The incidence and mortality rates of influenza and dengue have been high in most Asian countries. Vaccination is the one of the most important and successful public health measures in controlling many infectious diseases such as poliomyelitis, measles, and Japanese encephalitis. However, the etiological agents of these two diseases are RNA viruses involving different subtypes/serotypes with dynamic changes each year, posing greater challenges in future years. Methods and materials: Field epidemiology during outbreak periods provides the best chance in fully understanding dynamic changes of the studied virus and evaluating the impact of mismatched human influenza viruses. On the other hand, incidence and hospitalization rates, case fatality rates and seroepidemiological study of dengue are important in assessing public health needs and community effectiveness of vaccine as well as in finding target population to receive dengue vaccine. Results: We found that the higher isolation rate of seasonal influenza A (H3N2) and the circulating vaccine-mismatched human influenza viruses, the greater monthly P&I mortality in elderly, using a negative binomial model. Moreover, those years with high excess P&I mortality had the circulating virus strains with a lower amino acid identity percentages of HA1 protein with the vaccine strains. During 2009 pandemic influenza, one unique adaptive mutant at E374K of HA2 in A (H1N1/09) viruses evolved through the epidemic, emerged as a dominant strain at post-peak period, extended in areas with high population densities before mass immunization, persisted till ten months post-nationwide vaccination, and were finally fixed as herd immunity developed. Recently, different tetravalent dengue vaccines providing immunity against all four serotypes of dengue virus (DENV) have been developed to avoid antibody-dependent enhancement (ADE) from secondary DENV infection. The initial clinical trials in many cities of dengue-endemic countries did reduce the incidence and hospitalization rates of dengue. Conclusion: Weekly monitoring excess P&I mortality and evaluating laboratory-confirmed human influenza cases with/without receiving flu vaccines plus serological tests on antigenic variation can assure public health effectiveness of influenza vaccines. Once novel influenza virus emerges with worldwide spread, virological surveillance in high population-density areas can assess the virus with high epidemic potential. Furthermore, more immuno-epidemiological studies are needed for dengue in different settings.

During major epidemic outbreaks, demand for healthcare workers (HCWs) grows even as the extreme pressures they face cause declining availability. We draw on Taiwan's severe acute respiratory syndrome (SARS) experience to argue that a modified form of traffic control bundling (TCB) protects HCW safety and by extension strengthens overall coronavirus disease 2019 (COVID-19) epidemic control.

Background: Dengue epidemics in southern Taiwan have been documented for decades and most outbreaks have been triggered by imported cases from Southeast Asia. Since 1998, most local dengue outbreaks have happened in different magnitudes and scales in southern Taiwan, with the two greatest numbers of cases occurring in Kaohsiung City and Tainan City. From 2014 to 2015, there had been two largest outbreaks caused by DENV-1 and DENV-2 in two consecutive years in Kaohsiung City, while Tainan City had encountered the most severe outbreak of DENV-2 in 2015. Methods and materials:: To understand the post-outbreak herd immunity against dengue virus (DENV) infection, we investigated the seroprevalence rates of DENV infection from 5253 serum samples of older residents in Kaohsiung and Tainan Cities. Seropositive samples were verified to exclude Japanese encephalitis virus (JEV) cross-reaction and the DENV-reconfirmed samples were used to identify viral serotypes and secondary infection. Results: The overall seroprevalence rates of DENV infection was 19.27%, from 5.88% among < 50-year-old to 52.32% in age groups > 80 years. Both DENV seroprevalence rates in Kaohsiung and Tainan increased with age and were significantly lower below 50 years of age than those 70 to 79 and > 80 age groups. Additionally, DENV-1, DENV-2, and DENV-3 covered wider age groups in Kaohsiung, whereas DENV-2 was dominant in Tainan. Conclusion: Both the overall and age-specific seroprevalence rates in southern Taiwan were much lower than those in many Southeast Asian countries. Besides, the predominant DENV serotype in each of the large-scale past epidemics of dengue in Taiwan is also different from multiple DENV serotypes have been co-circulating in many dengue-endemic countries in Southeast Asia. Therefore, our seroepidemiological results plus the predominant DENV serotype causing major outbreaks verify that dengue in Taiwan has not been endemic yet.

Background: Taiwan belongs to both tropical and subtropical climate zones, which results in distinctions of spatial distribution between vectors and cases of human infection of dengue. The vectors of dengue in Taiwan comprise Aedes albopictus (distributed island-wide) and Aedes aegypti (located in southern Taiwan). The aims of this study were to (1) find out similarities and differences in epidemiologic characteristics between indigenous dengue cases in southern and other regions of Taiwan, and (2) compare case fatality rates and incidence rates of dengue cases in outbreak cities in southern and non-southern Taiwan. Methods and materials: The data was obtained from Taiwan centers for disease control. Demographic data was from statistics of Ministry of Interior. Epidemiologic characteristics, case fatality rates and incidence rates of 733 indigenous dengue cases from Aedes albopictus prevalent regions (Taipei, New Taipei, Taichung, and Taoyuan Cities) and 74309 indigenous dengue cases from southern Taiwan (Kaohsiung, Tainan, and Pingtung Cities) during 1998–2019 were analyzed by SAS 9.4 and QGIS 3.8. Results: Dengue clusters were frequently detected in southern Taiwan, while cases in other regions were more likely to be sporadic. Age distribution for cases in southern Taiwan and other regions peaked at 50–59 years (N = 13576, 18.27%) and 30–39 years (N = 159, 21.69%), respectively with statistical significance (p < 0.001). The percentage of male dengue patients in southern Taiwan was significantly lower than those in other regions [N = 36685 (N = 49.37%) vs. N = 403 (54.98%)] (p < 0.0025). Case fatality rates (CFRs) of dengue in southern Taiwan in 2002, 2014, 2015, and 2016 were 0.36%, 0.14%, 0.53%, 1.30%, much higher than CFR in Aedes albopictus prevalent area. Incidence rate of indigenous cases per 100 thousand population in southern Taiwan in 2002, 2014, 2015, and 2016 (95.57, 279.03, 778.90, 6.78) were much higher than those in other regions (0.26, 0.43, 2.55, 0.05). Conclusion: Dengue in Aedes aegypti-affected areas resulted in larger-scale and more severe epidemics involving more numbers of clusters than in areas with only Aedes albopictus. Therefore, more active surveillance and early detection of dengue cases in Aedes aegypti distributed areas is needed.

In recent decades, the global incidence of dengue has increased. Affected countries have responded with more effective surveillance strategies to detect outbreaks early, monitor the trends, and implement prevention and control measures. We have applied newly developed machine learning approaches to identify laboratory-confirmed dengue cases from 4,894 emergency department patients with dengue-like illness (DLI) who received laboratory tests. Among them, 60.11% (2942 cases) were confirmed to have dengue. Using just four input variables [age, body temperature, white blood cells counts (WBCs) and platelets], not only the state-of-the-art deep neural network (DNN) prediction models but also the conventional decision tree (DT) and logistic regression (LR) models delivered performances with receiver operating characteristic (ROC) curves areas under curves (AUCs) of the ranging from 83.75% to 85.87% [for DT, DNN and LR: 84.60% ± 0.03%, 85.87% ± 0.54%, 83.75% ± 0.17%, respectively]. Subgroup analyses found all the models were very sensitive particularly in the pre-epidemic period. Pre-peak sensitivities (<35 weeks) were 92.6%, 92.9%, and 93.1% in DT, DNN, and LR respectively. Adjusted odds ratios examined with LR for low WBCs [≤ 3.2 (x103/μL)], fever (≥38°C), low platelet counts [< 100 (x103/μL)], and elderly (≥ 65 years) were 5.17 [95% confidence interval (CI): 3.96-6.76], 3.17 [95%CI: 2.74-3.66], 3.10 [95%CI: 2.44-3.94], and 1.77 [95%CI: 1.50-2.10], respectively. Our prediction models can readily be used in resource-poor countries where viral/serologic tests are inconvenient and can also be applied for real-time syndromic surveillance to monitor trends of dengue cases and even be integrated with mosquito/environment surveillance for early warning and immediate prevention/control measures. In other words, a local community hospital/clinic with an instrument of complete blood counts (including platelets) can provide a sentinel screening during outbreaks. In conclusion, the machine learning approach can facilitate medical and public health efforts to minimize the health threat of dengue epidemics. However, laboratory confirmation remains the primary goal of surveillance and outbreak investigation.

Dengue virus (DENV) infections may cause life-threatening dengue hemorrhagic fever (DHF). Suppressed protective immunity was shown in these patients. Although several hypotheses have been formulated, the mechanism of DENV-induced immunosuppression remains unclear. Previously, we found that cross-reactive antibodies against tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) receptor 1 (death receptor 4 [DR4]) were elicited in DHF patients, and that anti-DR4 autoantibody fractions were elicited by nonstructural protein 1 (NS1) immunizations in experimental mice. In this study, we found that anti-DR4 antibodies could suppress B lymphocyte function in vitro and in vivo. Treatment with the anti-DR4 immunoglobulin (Ig) induced caspase-dependent cell death in immortalized B lymphocyte Raji cells in vitro. Anti-DR4 Igs elicited by NS1 and DR4 immunizations markedly suppressed mouse spleen transitional T2 B (IgM+IgD+), bone marrow pre-pro-B (B220+CD43+), pre-B (B220+CD43−), and mature B cell (B220+IgD+) subsets in mice. Furthermore, functional analysis revealed that the pre-elicitation of anti-NS1 and anti-DR4 Ig titers suppressed subsequently neutralizing antibody production by immunization with DENV envelop protein. Our data suggest that the elicitation of anti-DR4 titers through DENV NS1 immunization plays a suppressive role in humoral immunity in mice.

Background: Large-scale outbreaks of dengue occurred in 2014–2015 in southern Taiwan. As the fatal dengue cases in Taiwan were the oldest in the world, it is important to think best way of syndromic surveillance in elderly to minimize case fatality rate. This study had 3 aims: (1) to find out the differences in most frequent syndromic groups between adults and elderly, (2) to compare most likely syndromic groups in elderly with and without comorbidities, and (3) to extend Taiwan's findings to elderly data in other countries. Methods and materials: All the data were obtained laboratory-confirmed dengue cases in Taiwan during years with large-scale epidemics Comorbidity data were extracted from hospital records. Other variables included age groups, gender, residential areas and important epidemiological data were obtained from the Centers for disease Control in Taiwan (Taiwan-CDC). Chi-square test was used to find differences in percentages of symptoms and signs. Results: In the 2015 outbreak covering 2358 patients revealed that the increasing mean age was associated with clinically severe group (Group A: 38.4 years, Group B: 62.7 years, Group C: 70.5 years) and more critical signs (i.e. Glasgow Coma Scale <8.), as well as higher 30-day case fatality rate (Group A: 0%, Group B: 0.5%, Group C: 46.2%). Importantly, the percentage of comorbidities of diabetes mellitus (DM), hypertension (HT), and end-stage renal disease (ERD) increased with clinical severity group as comorbidity rates increased with age elevated (DM: 4.2% in Group A, 28.5% in Group B, and jumping to 47.7% in Group C; HT: Group A, 7.1%, group B, 55.0%, Group C, 70.8%; ERD: Group A, 0.3%, Group B, 2.1%, Group C, 10.8%) Syndromic surveillance of dengue in elderly must consider comorbidities to detect early cases. In Singapore with 76.2% of Chinese population also found dengue patients older than 60 years had 2.75 times higher severe organ involvement (SOI) risk than younger adults (12–29 years). Moreover, patients who presented with co-morbidities of DM and HT had 1.63 times higher SOI risk than those without these co-morbidities. Conclusion: Syndromic surveillance of dengue in elderly must consider comorbidities to detect early cases and it can help surveillance of other infectious diseases in elderly.