Moving Biosurveillance Beyond Coded Data Using AI for Symptom Detection From Physician Notes: Retrospective Cohort Study.

BACKGROUND: Real-time surveillance of emerging infectious diseases necessitates a dynamically evolving, computable case definition, which frequently incorporates symptom-related criteria. For symptom detection, both population health monitoring platforms and research initiatives primarily depend on structured data extracted from electronic health records. OBJECTIVE: This study sought to validate and test an artificial intelligence (AI)-based natural language processing (NLP) pipeline for detecting COVID-19 symptoms from physician notes in pediatric patients. We specifically study patients presenting to the emergency department (ED) who can be sentinel cases in an outbreak. METHODS: Subjects in this retrospective cohort study are patients who are 21 years of age and younger, who presented to a pediatric ED at a large academic children's hospital between March 1, 2020, and May 31, 2022. The ED notes for all patients were processed with an NLP pipeline tuned to detect the mention of 11 COVID-19 symptoms based on Centers for Disease Control and Prevention (CDC) criteria. For a gold standard, 3 subject matter experts labeled 226 ED notes and had strong agreement (F1-score=0.986; positive predictive value [PPV]=0.972; and sensitivity=1.0). F1-score, PPV, and sensitivity were used to compare the performance of both NLP and the International Classification of Diseases, 10th Revision (ICD-10) coding to the gold standard chart review. As a formative use case, variations in symptom patterns were measured across SARS-CoV-2 variant eras. RESULTS: There were 85,678 ED encounters during the study period, including 4% (n=3420) with patients with COVID-19. NLP was more accurate at identifying encounters with patients that had any of the COVID-19 symptoms (F1-score=0.796) than ICD-10 codes (F1-score =0.451). NLP accuracy was higher for positive symptoms (sensitivity=0.930) than ICD-10 (sensitivity=0.300). However, ICD-10 accuracy was higher for negative symptoms (specificity=0.994) than NLP (specificity=0.917). Congestion or runny nose showed the highest accuracy difference (NLP: F1-score=0.828 and ICD-10: F1-score=0.042). For encounters with patients with COVID-19, prevalence estimates of each NLP symptom differed across variant eras. Patients with COVID-19 were more likely to have each NLP symptom detected than patients without this disease. Effect sizes (odds ratios) varied across pandemic eras. CONCLUSIONS: This study establishes the value of AI-based NLP as a highly effective tool for real-time COVID-19 symptom detection in pediatric patients, outperforming traditional ICD-10 methods. It also reveals the evolving nature of symptom prevalence across different virus variants, underscoring the need for dynamic, technology-driven approaches in infectious disease surveillance.

Response to Genovesi et al.: Ant biosurveillance should come before invasion.

Menchetti et al. respond to the letter of Genovesi et al. and contribute new records of the red imported fire ant in Sicily.

Comparison of analytical performance and economic value of two biosurveillance methods for tracking SARS-COV-2 variants of concern.

The development of biosurveillance programs with strong analytical performance and economically accessible protocols is essential for monitoring viral pathogens. Throughout the COVID-19 pandemic, whole-genome sequencing (WGS) has been the prevailing technology for SARS-CoV-2 variant of concern (VOC) detection. While WGS offers benefits, it is a lengthy process, financially and technically straining for scalable viral tracking. The aim of this study was to compare the analytical performance and economic feasibility of WGS and PCR mutation panels for distinguishing six known VOCs: Alpha (B.1.1.7 and Q.4), Gamma (P.1), Delta (B.1.617.2 and AY.4.2), and Omicron. (B.1.1.529.1). In all, 78 SARS-CoV-2-positive samples were collected from April to December 2021 at Northeastern University (Cabot Testing Site, Boston, MA, USA) for genotyping PCR and WGS analysis. MagMax Viral/Pathogen II Nucleic Acid Isolation and TaqPath COVID-19 Combo Kits were used for RNA extraction and SARS-CoV-2 confirmation. VOC discrimination was assessed using two TaqMan SARS-CoV-2 single nucleotide polymorphism (SNP) assay layouts, and Ion Torrent WGS. In November 2021, the mutation panel demonstrated marked versatility by detecting the emerging Omicron variant reported by South Africa. SNP panel analysis yielded the following 78 VOC identifications: Alpha B.1.1.7 (N = 20), Alpha Q.4 (N = 3), Gamma P.1 (N = 1), Delta B.1.617.2 (N = 30), Delta AY.4.2 (N = 3), and Omicron B.1.1.529.1 (N = 20) with one undetermined (N = 1) sample. Genotyping mutation panels designated lineages in 77 of 78 samples, 46/78 were confirmed by WGS, while 32 samples failed WGS lineage assignment. RT-PCR genotyping panels offer pronounced throughput and sensitivity and provide an economically advantageous technique for SARS-CoV-2 biosurveillance.IMPORTANCEThe results presented in our manuscript demonstrate how the value of simplistic and reliable molecular assays coupled with the core scientific principle of standardization can be overlooked by the charm of more sophisticated assays and instrumentation. This effect can often be amplified during tumultuous public health events, such as the COVID-19 pandemic. By adapting standardized PCR mutation panels to detect prominently circulating SARS-CoV-2 variants, we were able to better assess the potential health impacts of rising positivity rates and transmission clusters within the Northeastern University population. While several literature publications utilizing genotyping PCR and NGS have a similar scope to ours, many investigations lack sufficiently standardized genotyping PCR and NGS bioinformatics inclusionary/exclusionary criteria for SARS-CoV-2 variant identification. Finally, the economic benefits of standardized PCR mutation panels would allow for global implementation of biosurveillance, rather than reserving biosurveillance to more economically developed nations.

Portable RT-PCR and MinION Nanopore Sequencing as a Proof-of-Concept SARS-CoV-2 Biosurveillance in Wastewater.

In this study, wastewater samples collected from a participating sentinel site were initially screened for the presence or absence of SARS-CoV-2 RNA using portable RT-PCR, with positive samples sequenced using a handheld MinION nanopore sequencing device. Genomic biosurveillance of SARS-CoV-2 and its variants within wastewater has been established as an early warning system of infectious disease spread in a given catchment area, due to good correlation between spikes in viral levels detected in wastewater coincident with increases in COVID-19 incidence rates. Moreover, viral titers detected in a single wastewater sample are reflective of pre-symptomatic, asymptomatic, and post-symptomatic cases, making wastewater-based epidemiology (WBE) a cost-effective, non-invasive public health surveillance method complementary to clinical diagnostic testing. The results of this study demonstrate the utility of population-scale SARS-CoV-2 epidemiology for insights into the viral evolution and transmission dynamics associated with specific SARS-CoV-2 variants that are necessary for effective strategies of containment and timely deployment of appropriate countermeasures.

Unveiling invasive insect threats to plant biodiversity: Leveraging eDNA metabarcoding and saturated salt trap solutions for biosurveillance.

The negative global impacts of invasive alien species (IAS) on biodiversity are second only to habitat loss. eDNA metabarcoding allows for a faster and more comprehensive evaluation of community species composition, with a higher taxonomic resolution and less taxonomic expertise required than traditional morphological-based biosurveillance. These advantages have positioned eDNA metabarcoding as the standard method for molecular-based detection of invasive alien species, where fast and accurate detectability allows prompt responses to mitigate their adverse effects. Here, eDNA metabarcoding is used for biosurveillance of invasive alien species regulated by Canada in high-risk areas with four main objectives: i) validate the effectiveness of eDNA metabarcoding of salt trap solutions as a molecular technique for IAS detection, ii) compare detection from DNA extracts obtained from filter quarters versus whole filters, iii) benchmark two different bioinformatic pipelines (MetaWorks and mBRAVE), and iv) compare canopy and ground level trapping. eDNA from up to five IAS (Agrilus planipennis, Daktulosphaira vitifoliae, Lymantria dispar, Popillia japonica, and Trichoferus campestris) were successfully detected across years from 2017 to 2022 in southern Ontario, Canada, with successful morphological validation for all except Lymantria dispar and Trichoferus campestris. Analysis of filter quarters in contrast to whole filters was demonstrated to be insufficient for effective IAS detection in each sample. All IAS were detected in only one filter quarter, suggesting a patchy eDNA distribution on the filter. The MetaWorks and mBRAVE bioinformatics pipelines proved effective in identifying IAS, with MetaWorks yielding a higher success rate when comparing molecular and morphological identifications. Ground-level and canopy-level sampling showed differential IAS recovery rates based on the molecular detection, which also varied per collection year, with all found IAS detected at the canopy level in 2022 while only one (Lymantria dispar) in 2020. The present study ratifies the efficacy and importance of eDNA-based detection in a regulatory context and the utility of adding eDNA metabarcoding of saturated salt trap solutions, a critical tool for IAS detection.

Applied genomics for identification of virulent biothreats and for disease outbreak surveillance.

Fortifying our preparedness to cope with biological threats by identifying and targeting virulence factors may be a preventative strategy for curtailing infectious disease outbreak. Virulence factors evoke successful pathogenic invasion, and the science and technology of genomics offers a way of identifying them, their agents and evolutionary ancestor. Genomics offers the possibility of deciphering if the release of a pathogen was intentional or natural by observing sequence and annotated data of the causative agent, and evidence of genetic engineering such as cloned vectors at restriction sites. However, to leverage and maximise the application of genomics to strengthen global interception system for real-time biothreat diagnostics, a complete genomic library of pathogenic and non-pathogenic agents will create a robust reference assembly that can be used to screen, characterise, track and trace new and existing strains. Encouraging ethical research sequencing pathogens found in animals and the environment, as well as creating a global space for collaboration will lead to effective global regulation and biosurveillance.

Development of a Pan-Filoviridae SYBR Green qPCR Assay for Biosurveillance Studies in Bats.

Recent studies have indicated that bats are hosts to diverse filoviruses. Currently, no pan-filovirus molecular assays are available that have been evaluated for the detection of all mammalian filoviruses. In this study, a two-step pan-filovirus SYBR Green real-time PCR assay targeting the nucleoprotein gene was developed for filovirus surveillance in bats. Synthetic constructs were designed as representatives of nine filovirus species and used to evaluate the assay. This assay detected all synthetic constructs included with an analytical sensitivity of 3-31.7 copies/reaction and was evaluated against the field collected samples. The assay's performance was similar to a previously published probe based assay for detecting Ebola- and Marburgvirus. The developed pan-filovirus SYBR Green assay will allow for more affordable and sensitive detection of mammalian filoviruses in bat samples.

Analysis of metatranscriptomic methods to enable wastewater-based biosurveillance of all infectious diseases.

Introduction: Wastewater-based surveillance emerged during the COVID-19 pandemic as an efficient way to quickly screen large populations, monitor infectious disease transmission over time, and identify whether more virulent strains are becoming more prevalent in the region without burdening the health care system with individualized testing. Ohio was one of the first states to implement wastewater monitoring through its Ohio Coronavirus Wastewater Monitoring Network (OCWMN), originally tracking the prevalence of COVID-19 by quantitative qPCR from over 67 sites across the state. The OCWMN evolved along with the pandemic to include sequencing the SARS-CoV-2 genome to assess variants of concern circulating within the population. As the pandemic wanes, networks such as OCWMN can be expanded to monitor other infectious diseases and outbreaks of interest to the health department to reduce the burden of communicable diseases. However, most surveillance still utilizes qPCR based diagnostic tests for individual pathogens, which is hard to scale for surveillance of multiple pathogens. Methods: Here we have tested several genomic methods, both targeted and untargeted, for wastewater-based biosurveillance to find the most efficient procedure to detect and track trends in reportable infectious diseases and outbreaks of known pathogens as well as potentially novel pathogens or variants on the rise in our communities. RNA extracts from the OCWMN were provided weekly from 10 sites for 6 weeks. Total RNA was sequenced from the samples on the Illumina NextSeq and on the MinION to identify pathogens present. The MinION long read platform was also used to sequence SARS-CoV-2 with the goal of reducing the complexity of variant calling in mixed populations as occurs with short Illumina reads. Finally, a targeted hybridization approach was tested for compatibility with wastewater RNA samples. Results and discussion: The data analyzed here provides a baseline assessment that demonstrates that wastewater is a rich resource for infectious disease epidemiology and identifies technology gaps and potential solutions to enable this resource to be used by public health laboratories to monitor the infectious disease landscape of the regions they serve.

Required Elements for an Integrated Biosurveillance Platform: A Capacity and Needs Assessment of the Central Government in the Republic of Korea.

New and reemerging infectious disease outbreaks threaten human safety worldwide, increasing the urgency to implement biosurveillance systems that enhance government capacity in public health emergency preparedness and response. To do so, it is necessary to evaluate existing surveillance and response activities and identify potential barriers at the national level. This study aimed to assess the current status and readiness of government agencies in South Korea, particularly for information sharing and use, and to identify barriers and opportunities in developing an agency-integrated biosurveillance system. The target sample size was 66 government officials, working at 6 relevant government ministries. We invited a total of 100 officials to participate. A total of 34 government officials completed the survey (34.0% response rate), 18 (52.9%) of whom were affiliated with the Korea Disease Control and Prevention Agency or the Ministry of Health and Welfare. Findings revealed that information sharing between government agencies occurred frequently, but a discrepancy existed in terms of the type of information shared and stored. Although information sharing with other agencies and ministries occurred at all stages-prevention, preparation, response, and recovery-it mostly revolved around preventive activities, with no respondents reportedly sharing recovery-related information. An agency-integrated biosurveillance system is crucial in preparing for the next pandemic, as well as supporting information sharing, analysis, and interpretation across humans, animals, and the environment. It is key to national and global health security.

Artificial intelligence in public health: the potential of epidemic early warning systems.

The use of artificial intelligence (AI) to generate automated early warnings in epidemic surveillance by harnessing vast open-source data with minimal human intervention has the potential to be both revolutionary and highly sustainable. AI can overcome the challenges faced by weak health systems by detecting epidemic signals much earlier than traditional surveillance. AI-based digital surveillance is an adjunct to-not a replacement of-traditional surveillance and can trigger early investigation, diagnostics and responses at the regional level. This narrative review focuses on the role of AI in epidemic surveillance and summarises several current epidemic intelligence systems including ProMED-mail, HealthMap, Epidemic Intelligence from Open Sources, BlueDot, Metabiota, the Global Biosurveillance Portal, Epitweetr and EPIWATCH. Not all of these systems are AI-based, and some are only accessible to paid users. Most systems have large volumes of unfiltered data; only a few can sort and filter data to provide users with curated intelligence. However, uptake of these systems by public health authorities, who have been slower to embrace AI than their clinical counterparts, is low. The widespread adoption of digital open-source surveillance and AI technology is needed for the prevention of serious epidemics.

Can Water-Only DNA Extraction Reduce the Logistical Footprint of Biosurveillance and Planetary Health Diagnostics? Toward a New Method.

The coronavirus disease-2019 (COVID-19) pandemic has raised the stakes for planetary health diagnostics. Because pandemics pose enormous burdens on biosurveillance and diagnostics, reduction of the logistical burdens of pandemics and ecological crises is essential. Moreover, the disruptive effects of catastrophic bioevents impact the supply chains in both highly populated urban centers and rural communities. One "upstream" focus of methodological innovation in biosurveillance is the footprint of Nucleic Acid Amplification Test (NAAT)-based assays. We report in this study a water-only DNA extraction, as an initial step in developing future protocols that may require few expendables, and with low environmental footprints, in terms of wet and solid laboratory waste. In the present work, boiling-hot distilled water was used as the main cell lysis agent for direct polymerase chain reactions (PCRs) on crude extracts. After evaluation (1) in blood and mouth swabs for human biomarker genotyping, and (2) in mouth swabs and plant tissue for generic bacterial or fungal detection, and using different combinations of extraction volume, mechanical assistance, and extract dilution, we found the method to be applicable in low-complexity samples, but not in high-complexity ones such as blood and plant tissue. In conclusion, this study examined the doability of a lean approach for template extraction in the case of NAAT-based diagnostics. Testing our approach with different biosamples, PCR settings, and instruments, including portable ones for COVID-19 or dispersed applications, warrant further research. Minimal resources analysis is a concept and practice, vital and timely for biosurveillance, integrative biology, and planetary health in the 21st century.

Development of an array of molecular tools for the identification of khapra beetle (Trogoderma granarium), a destructive beetle of stored food products.

Trogoderma granarium Everts, the khapra beetle, native to the Indian subcontinent, is one of the world's most destructive pests of stored food products. Early detection of this pest facilitates prompt response towards the invasion and prevents the need for costly eradication efforts. Such detection requires proper identification of T. granarium, which morphologically resembles some more frequently encountered, non-quarantine congeners. All life stages of these species are difficult to distinguish using morphological characters. Additionally, biosurveillance trapping can result in the capture of large numbers of specimens awaiting identification. To address these issues, we aim to develop an array of molecular tools to rapidly and accurately identify T. granarium among non-target species. Our crude, cheap DNA extraction method performed well for Trogoderma spp. and is suitable for downstream analyses including sequencing and real-time PCR (qPCR). We developed a simple quick assay usingrestriction fragment length polymorphism to distinguish between T. granarium and the closely related, congeneric T. variabile Ballion and T. inclusum LeConte. Based on newly generated and published mitochondrial sequence data, we developed a new multiplex TaqMan qPCR assay for T. granarium with improved efficiency and sensitivity over existing qPCR assays. These new tools benefit regulatory agencies and the stored food products industry by providing cost- and time-effective solutions to enhance the identification of T. granarium from related species. They can be added to the existing pest detection toolbox. The selection of which method to use would depend on the intended application.

Managing the Transition to Widespread Metagenomic Monitoring: Policy Considerations for Future Biosurveillance.

The technological possibilities and future public health importance of metagenomic sequencing have received extensive attention, but there has been little discussion about the policy and regulatory issues that need to be addressed if metagenomic sequencing is adopted as a key technology for biosurveillance. In this article, we introduce metagenomic monitoring as a possible path to eventually replacing current infectious disease monitoring models. Many key enablers are technological, whereas others are not. We therefore highlight key policy challenges and implementation questions that need to be addressed for "widespread metagenomic monitoring" to be possible. Policymakers must address pitfalls like fragmentation of the technological base, private capture of benefits, privacy concerns, the usefulness of the system during nonpandemic times, and how the future systems will enable better response. If these challenges are addressed, the technological and public health promise of metagenomic sequencing can be realized.

Genomic biosurveillance detects a sexual hybrid in the sudden oak death pathogen.

Invasive exotic pathogens pose a threat to trees and forest ecosystems worldwide, hampering the provision of essential ecosystem services such as carbon sequestration and water purification. Hybridization is a major evolutionary force that can drive the emergence of pathogens. Phytophthora ramorum, an emergent pathogen that causes the sudden oak and larch death, spreads as reproductively isolated divergent clonal lineages. We use a genomic biosurveillance approach by sequencing genomes of P. ramorum from survey and inspection samples and report the discovery of variants of P. ramorum that are the result of hybridization via sexual recombination between North American and European lineages. We show that these hybrids are viable, can infect a host and produce spores for long-term survival and propagation. Genome sequencing revealed genotypic combinations at 54,515 single nucleotide polymorphism loci not present in parental lineages. More than 6,000 of those genotypes are predicted to have a functional impact in genes associated with host infection, including effectors, carbohydrate-active enzymes and proteases. We also observed post-meiotic mitotic recombination that could generate additional genotypic and phenotypic variation and contribute to homoploid hybrid speciation. Our study highlights the importance of plant pathogen biosurveillance to detect variants, including hybrids, and inform management and control.

Insights from biosurveillance: non-fatal opioid overdoses in Rhode Island 2019-21.

BACKGROUND AND AIMS: Opioids biosurveillance is a new approach to public health surveillance of non-fatal overdoses that relies upon laboratory analysis of residual biospecimens from hospitals treating opioids overdoses. In Rhode Island (RI), USA, hospitals report suspected opioid overdoses to the Department of Health. Residual specimens associated with these overdoses are submitted to the State Health Laboratories for further characterization. This surveillance project aimed to characterize non-fatal overdoses through toxicological testing of urine specimens associated with non-fatal overdoses during the initial 2-year period of biosurveillance implementation in RI to assess the feasibility and public health utility of this approach. METHODS: This study included individuals who presented for treatment for a suspected opioid overdose in 10 RI hospitals between July 2019 and June 2021. Urine samples were received for 1354 unique overdose encounters corresponding to reported overdoses. Some individuals experienced multiple overdoses during this time. Urine samples were extracted and then analyzed by liquid chromatography tandem mass spectrometry with a panel consisting of 1033 opiates, synthetic opioids, fentanyl analogs and select metabolites. Temporal and spatial trends were evaluated for the studied population. RESULTS: A total of 1354 samples were tested for the presence of opioids in urine collected from individuals who experienced a suspected overdose. Fentanyl (and/or norfentanyl) was present in 79% of all samples in which opioids were found (n = 1033). Fentanyl analogs varied in their contribution to these totals, with observations ranging as high as 35% of all opioid-containing samples in August 2019 and May 2021. CONCLUSIONS: Laboratory identification of opioids involved in suspected overdoses shows that fentanyl and its analogs are the main drivers of the opioids overdose epidemic in Rhode Island, USA. The biosurveillance approach is unique in its ability to quantify contributions of novel fentanyl analogs to the burden of overdoses.

Biosurveillance of Drug Overdoses and Substance Misuse Treated in Selected Emergency Departments in Minnesota, 2017-2020.

OBJECTIVES: Increasing knowledge about the toxicology of drug overdose and substance misuse (DOSM) is important in improving our understanding of the epidemic. We describe the Minnesota Drug Overdose and Substance Use Pilot Surveillance Activity, which started collecting data on emergency department (ED) visits attributable to DOSM in 2017, with a focus on the toxicology results of a subset of clinical encounters. METHODS: From November 1, 2017, through January 30, 2020, we collected near-real-time data on DOSM-related ED encounters. The Minnesota Department of Health Public Health Laboratory tested leftover clinical specimens (blood and/or urine) for the presence of various substances for patients who died, were hospitalized, had an atypical clinical presentation, or were part of a local drug overdose cluster. Testing looked for >250 drugs or their metabolites, including those commonly misused (eg, methamphetamine, cocaine), prescription medications, synthetic cannabinoids and cathinones, and opioids. We describe characteristics of the overall group and a subgroup of clinical encounters with toxicology results. RESULTS: Specimens submitted from 6 EDs during the study period represented 239 clinical encounters. Methamphetamine was the most frequently detected substance (67.4%) but was suspected in only 45.6% of encounters. At least 1 opioid was detected in 42.5% of encounters but suspected in only 29.7%. Testing also detected potential adulterants and additives (eg, fentanyl, fentanyl analogues, levamisole) and showed frequent patient exposure to substances not reported by patients or suspected by clinicians. Nearly half (44.4%) of clinical encounters had >1 substance detected. CONCLUSIONS: ED surveillance for DOSM encounters, enhanced by toxicology testing, can provide local situational awareness on overdoses, prevent potential mischaracterization of the true drug overdose epidemic, and inform harm reduction and drug overdose prevention efforts.

Known and Unknown Transboundary Infectious Diseases as Hybrid Threats.

The pathogenicity, transmissibility, environmental stability, and potential for genetic manipulation make microbes hybrid threats that could blur the distinction between peace and war. These agents can fall below the detection, attribution, and response capabilities of a nation and seriously affect their health, trade, and security. A framework that could enhance horizon scanning regarding the potential risk of microbes used as hybrid threats requires not only accurately discriminating known and unknown pathogens but building novel scenarios to deploy mitigation strategies. This demands the transition of analyst-based biosurveillance tracking a narrow set of pathogens toward an autonomous biosurveillance enterprise capable of processing vast data streams beyond human cognitive capabilities. Autonomous surveillance systems must gather, integrate, analyze, and visualize billions of data points from different and unrelated sources. Machine learning and artificial intelligence algorithms can contextualize capability information for different stakeholders at different levels of resolution: strategic and tactical. This document provides a discussion of the use of microorganisms as hybrid threats and considerations to quantitatively estimate their risk to ensure societal awareness, preparedness, mitigation, and resilience.