Correlation-Based Discovery of Disease Patterns for Syndromic Surveillance

Early outbreak detection is a key aspect in the containment of infectious diseases, as it enables the identification and isolation of infected individuals before the disease can spread to a larger population. Instead of detecting unexpected increases of infections by monitoring confirmed cases, syndromic surveillance aims at the detection of cases with early symptoms, which allows a more timely disclosure of outbreaks. However, the definition of these disease patterns is often challenging, as early symptoms are usually shared among many diseases and a particular disease can have several clinical pictures in the early phase of an infection. As a first step toward the goal to support epidemiologists in the process of defining reliable disease patterns, we present a novel, data-driven approach to discover such patterns in historic data. The key idea is to take into account the correlation between indicators in a health-related data source and the reported number of infections in the respective geographic region. In an preliminary experimental study, we use data from several emergency departments to discover disease patterns for three infectious diseases. Our results show the potential of the proposed approach to find patterns that correlate with the reported infections and to identify indicators that are related to the respective diseases. It also motivates the need for additional measures to overcome practical limitations, such as the requirement to deal with noisy and unbalanced data, and demonstrates the importance of incorporating feedback of domain experts into the learning procedure.

Global variation in event-based surveillance for disease outbreak detection: A time series analysis (Preprint)

BACKGROUND Robust and flexible infectious disease surveillance is crucial for public health. Event-based surveillance (EBS) was developed to allow timely detection of infectious disease outbreaks using mostly web-based data. Despite its widespread use, EBS has not been evaluated systematically on a global scale in terms of outbreak detection performance. OBJECTIVE To assess the variation in timing and frequency of EBS reports compared to true outbreaks and to identify the determinants of variability, using the example of seasonal influenza epidemics in 24 countries. METHODS We obtained influenza-related reports from two EBS systems, HealthMap and the WHO Epidemic Intelligence from Open Sources (EIOS), and weekly virologic influenza counts from FluNet as a gold standard. Epidemic influenza periods were detected based on report frequency using Bayesian change point analysis. Timely sensitivity, i.e., outbreak detection within the first two weeks before or after an outbreak onset, was calculated along with sensitivity, specificity, positive predictive value, and timeliness of detection. Linear regressions were performed to assess the influence of country-specific factors on EBS performance. RESULTS Overall, monitoring the frequency of EBS reports detected 73.5% of outbreaks, but only 9.2% within two weeks of onset; in the best case, monitoring the frequency of health-related reports identified 29% of outbreaks within two weeks of onset. We observed a large degree of variability in all evaluation metrics across countries. The number of EBS reports available within a country, the human development index, and the country’s geographical location partially explained this variability. CONCLUSIONS Monitoring the frequency of EBS reports allowed just under 10% of seasonal influenza outbreaks to be detected in a timely manner in a worldwide analysis, with a large variability in detection capabilities. This article documents the global variation of EBS performance and demonstrates that monitoring report frequency alone in EBS may be insufficient for timely detection of outbreaks. Moreover, factors such as human development index and geographical location of a country may influence the performance of EBS and should be considered in future evaluations.

Signatures of Cholera Outbreak during the Yemeni Civil War, 2016–2019

The Global Task Force on Cholera Control (GTFCC) created a strategy for early outbreak detection, hotspot identification, and resource mobilization coordination in response to the Yemeni cholera epidemic. This strategy requires a systematic approach for defining and classifying outbreak signatures, or the profile of an epidemic curve and its features. We used publicly available data to quantify outbreak features of the ongoing cholera epidemic in Yemen and clustered governorates using an adaptive time series methodology. We characterized outbreak signatures and identified clusters using a weekly time series of cholera rates in 20 Yemeni governorates and nationally from 4 September 2016 through 29 December 2019 as reported by the World Health Organization (WHO). We quantified critical points and periods using Kolmogorov–Zurbenko adaptive filter methodology. We assigned governorates into six clusters sharing similar outbreak signatures, according to similarities in critical points, critical periods, and the magnitude of peak rates. We identified four national outbreak waves beginning on 12 September 2016, 6 March 2017, 28 May 2018, and 28 January 2019. Among six identified clusters, we classified a core regional hotspot in Sana’a, Sana’a City, and Al-Hudaydah—the expected origin of the national outbreak. The five additional clusters differed in Wave 2 and Wave 3 peak frequency, timing, magnitude, and geographic location. As of 29 December 2019, no governorates had returned to pre-Wave 1 levels. The detected similarity in outbreak signatures suggests potentially shared environmental and human-made drivers of infection; the heterogeneity in outbreak signatures implies the potential traveling waves outwards from the core regional hotspot that could be governed by factors that deserve further investigation.

A deep learning approach to real-time HIV outbreak detection using genetic data

Pathogen genomic sequence data are increasingly made available for epidemiological monitoring. A main interest is to identify and assess the potential of infectious disease outbreaks. While popular methods to analyze sequence data often involve phylogenetic tree inference, they are vulnerable to errors from recombination and impose a high computational cost, making it difficult to obtain real-time results when the number of sequences is in or above the thousands. Here, we propose an alternative strategy to outbreak detection using genomic data based on deep learning methods developed for image classification. The key idea is to use a pairwise genetic distance matrix calculated from viral sequences as an image, and develop convolutional neutral network (CNN) models to classify areas of the images that show signatures of active outbreak, leading to identification of subsets of sequences taken from an active outbreak. We showed that our method is efficient in finding HIV-1 outbreaks with R0>3, and overall a specificity exceeding 85% and sensitivity better than 70%. We validated our approach using data from HIV-1 CRF01 in Europe, containing both endemic sequences and a well-known dual outbreak in intravenous drug users. Our model accurately identified known outbreak sequences in the background of slower spreading HIV. Importantly, we detected both outbreaks early on, before they were over, implying that had this method been applied in real-time as data became available, one would have been able to intervene and possibly prevent the extent of these outbreaks. This approach is scalable to processing hundreds of thousands of sequences, making it useful for current and future real-time epidemiological investigations, including public health monitoring using large databases and especially for rapid outbreak identification.

Opening up safely: public health system requirements for ongoing COVID-19 management based on evaluation of Australia's surveillance system performance

Background Ongoing management of COVID-19 requires an evidence-based understanding of the performance of public health measures to date, and application of this evidence to evolving response objectives. This paper aims to define system requirements for COVID-19 management under future transmission and response scenarios, based on surveillance system performance to date. Methods From 1st November 2020 to 30th June 2021 community transmission was eliminated in Australia, allowing investigation of system performance in detecting novel outbreaks, including against variants of concern (VoCs). We characterised surveillance systems in place from peer-reviewed and publicly available data, analysed the epidemiological characteristics of novel outbreaks over this period, and assessed surveillance system sensitivity and timeliness in outbreak detection. These findings were integrated with analysis of other critical COVID-19 public health measures to establish requirements for future COVID-19 management. Findings Australia reported 25 epidemiologically distinct outbreaks and 5 distinct clusters of cases in the study period, all linked through genomic sequencing to breaches in quarantine facilities housing international travellers. Most (21/30, 70%) were detected through testing of those with acute respiratory illness in the community, and 9 through quarantine screening. For the 21 detected in the community, the testing rate (percent of the total State population tested in the week preceding detection) was 2.07% on average, was higher for those detected while prior outbreaks were ongoing. For 17/30 with data, the delay from the primary case to detection of the index case was, on average 4.9 days, with 10 of the 17 outbreaks detected within 5 days and 3 detected after > 7days. One outbreak was preceded by an unexpected positive wastewater detection. Of the 24 outbreaks in 2021, 20 had publicly available sequencing data, all of which were VoCs. Surveillance for future VoCs using a similar strategy to that used for detecting SARS-CoV-2 to date would necessitate a 100-1,000-fold increase in capacity for genomic sequencing. Interpretation Australia's surveillance systems performed well in detecting novel introduction of SARS- CoV-2 in a period when community transmission was eliminated, introductions were infrequent and case numbers were low. Detection relied on community surveillance in symptomatic members of the general population and quarantine screening, supported by comprehensive genomic sequencing. Once vaccine coverage is maximised, the priority for future COVID-19 control will shift to detection of SARS-CoV-2 VoCs associated with increased severity of disease in the vaccinated and vaccine ineligible. This will require ongoing investment in maintaining surveillance systems and testing of all international arrivals, alongside greatly increased genomic sequencing capacity. Other essential requirements for managing VoCs are maintaining outbreak response capacity and developing capacity to rapidly engineer, manufacture, and distribute variant vaccines at scale. The most important factor in management of COVID-19 now and into the future will continue to be how effectively governments support all sectors of the community to engage in control measures.

Increasing risk of breakthrough COVID-19 in outbreaks with high attack rates in European long-term care facilities, July to October 2021

We collected data from 10 EU/EEA countries on 240 COVID-19 outbreaks occurring from July−October 2021 in long-term care facilities with high vaccination coverage. Among 17,268 residents, 3,832 (22.2%) COVID-19 cases were reported. Median attack rate was 18.9% (country range: 2.8–52.4%), 17.4% of cases were hospitalised, 10.2% died. In fully vaccinated residents, adjusted relative risk for COVID-19 increased with outbreak attack rate. Findings highlight the importance of early outbreak detection and rapid containment through effective infection prevention and control measures.

ECTyper: in silico Escherichia coli serotype and species prediction from raw and assembled whole-genome sequence data

Escherichia coli is a priority foodborne pathogen of public health concern and phenotypic serotyping provides critical information for surveillance and outbreak detection activities. Public health and food safety laboratories are increasingly adopting whole-genome sequencing (WGS) for characterizing pathogens, but it is imperative to maintain serotype designations in order to minimize disruptions to existing public health workflows. Multiple in silico tools have been developed for predicting serotypes from WGS data, including SRST2, SerotypeFinder and EToKi EBEis, but these tools were not designed with the specific requirements of diagnostic laboratories, which include: speciation, input data flexibility (fasta/fastq), quality control information and easily interpretable results. To address these specific requirements, we developed ECTyper (https://github.com/phac-nml/ecoli_serotyping) for performing both speciation within Escherichia and Shigella , and in silico serotype prediction. We compared the serotype prediction performance of each tool on a newly sequenced panel of 185 isolates with confirmed phenotypic serotype information. We found that all tools were highly concordant, with 92–97 % for O-antigens and 98–100 % for H-antigens, and ECTyper having the highest rate of concordance. We extended the benchmarking to a large panel of 6954 publicly available E. coli genomes to assess the performance of the tools on a more diverse dataset. On the public data, there was a considerable drop in concordance, with 75–91 % for O-antigens and 62–90 % for H-antigens, and ECTyper and SerotypeFinder being the most concordant. This study highlights that in silico predictions show high concordance with phenotypic serotyping results, but there are notable differences in tool performance. ECTyper provides highly accurate and sensitive in silico serotype predictions, in addition to speciation, and is designed to be easily incorporated into bioinformatic workflows.

Acute Infectious Gastroenteritis: The Causative Agents, Omics-Based Detection of Antigens and Novel Biomarkers

Acute infectious gastroenteritis (AGE) is among the leading causes of mortality in children less than 5 years of age worldwide. There are many causative agents that lead to this infection, with rotavirus being the commonest pathogen in the past decade. However, this trend is now being progressively replaced by another agent, which is the norovirus. Apart from the viruses, bacteria such as Salmonella and Escherichia coli and parasites such as Entamoeba histolytica also contribute to AGE. These agents can be recognised by their respective biological markers, which are mainly the specific antigens or genes to determine the causative pathogen. In conjunction to that, omics technologies are currently providing crucial insights into the diagnosis of acute infectious gastroenteritis at the molecular level. Recent advancement in omics technologies could be an important tool to further elucidate the potential causative agents for AGE. This review will explore the current available biomarkers and antigens available for the diagnosis and management of the different causative agents of AGE. Despite the high-priced multi-omics approaches, the idea for utilization of these technologies is to allow more robust discovery of novel antigens and biomarkers related to management AGE, which eventually can be developed using easier and cheaper detection methods for future clinical setting. Thus, prediction of prognosis, virulence and drug susceptibility for active infections can be obtained. Case management, risk prediction for hospital-acquired infections, outbreak detection, and antimicrobial accountability are aimed for further improvement by integrating these capabilities into a new clinical workflow.

Botulism outbreak in a rural Ethiopia: a case series

Background Foodborne botulism, a toxin-mediated illness caused by Clostridium botulinum, is a public health emergency. Types A, B, and E C. botulinum toxins commonly cause human disease. Outbreaks are often associated with homemade and fermented foods. Botulism is rarely reported in Africa and has never been reported in Ethiopia. Case presentation In March 2015, a cluster of family members from the Wollega, Oromia region, western Ethiopia presented with a symptom constellation suggestive of probable botulism. Clinical examination, epidemiologic investigation, and subsequent laboratory work identified the cause of the outbreak to be accidental ingestion of botulinum toxin in a traditional chili condiment called “Kochi-kocha,” cheese, and clarified butter. Ten out of the fourteen family members who consumed the contaminated products had botulism (attack rate 71.4%) and five died (case fatality rate of 50%). Three of the patients were hospitalized, they presented with altered mental status (n = 2), profound neck and truncal weakness (n = 3), and intact extremity strength despite hyporeflexia (n = 3). The remnant food sample showed botulinum toxin type A with mouse bioassay and C. botulinum type A with culture. Blood drawn on day three of illness from 2/3 (66%) cases was positive for botulinum toxin type-A. Additionally, one of these two cases also had C. botulinum type A cultured from a stool specimen. Two of the cases received Botulism antitoxin (BAT). Conclusion These are the first confirmed cases of botulism in Ethiopia. The disease occurred due to the consumption of commonly consumed homemade foods. Definite diagnoses of botulism cases are challenging, and detailed epidemiologic and laboratory investigations were critical to the identification of this case series. Improved awareness of botulism risk and improved food preparation and storage may prevent future illnesses. The mortality rate of botulism in resource-limited settings remains high. Countries should make a concerted effort to stockpile antitoxin as that is the easiest and quickest intervention after outbreak detection.