Assessing 3 Outbreak Detection Algorithms in an Electronic Syndromic Surveillance System in a Resource-Limited Setting.
Ontology highlight
ABSTRACT:
SUBMITTER: Alsentzer E
PROVIDER: S-EPMC7454051 | biostudies-literature | 2020 Sep
REPOSITORIES: biostudies-literature
Publications
Assessing 3 Outbreak Detection Algorithms in an Electronic Syndromic Surveillance System in a Resource-Limited Setting.
Alsentzer Emily E Ballard Sarah-Blythe SB Neyra Joan J Vera Delphis M DM Osorio Victor B VB Quispe Jose J Blazes David L DL Loayza Luis L
Emerging infectious diseases 20200901 9
We evaluated the performance of X-bar chart, exponentially weighted moving average, and C3 cumulative sums aberration detection algorithms for acute diarrheal disease syndromic surveillance at naval sites in Peru during 2007-2011. The 3 algorithms' detection sensitivity was 100%, specificity was 97%-99%, and positive predictive value was 27%-46%. ...[more]
Similar Datasets
Project description:MOTIVATION:Public health authorities can provide more effective and timely interventions to protect populations during health events if they have effective multi-purpose surveillance systems. These systems rely on aberration detection algorithms to identify potential threats within large datasets. Ensuring the algorithms are sensitive, specific and timely is crucial for protecting public health. Here, we evaluate the performance of three detection algorithms extensively used for syndromic surveillance: the 'rising activity, multilevel mixed effects, indicator emphasis' (RAMMIE) method and the improved quasi-Poisson regression-based method known as 'Farrington Flexible' both currently used at Public Health England, and the 'Early Aberration Reporting System' (EARS) method used at the US Centre for Disease Control and Prevention. We model the wide range of data structures encountered within the daily syndromic surveillance systems used by PHE. We undertake extensive simulations to identify which algorithms work best across different types of syndromes and different outbreak sizes. We evaluate RAMMIE for the first time since its introduction. Performance metrics were computed and compared in the presence of a range of simulated outbreak types that were added to baseline data. RESULTS:We conclude that amongst the algorithm variants that have a high specificity (i.e. >90%), Farrington Flexible has the highest sensitivity and specificity, whereas RAMMIE has the highest probability of outbreak detection and is the most timely, typically detecting outbreaks 2-3 days earlier. AVAILABILITY AND IMPLEMENTATION:R codes developed for this project are available through https://github.com/FelipeJColon/AlgorithmComparison. SUPPLEMENTARY INFORMATION:Supplementary data are available at Bioinformatics online.
| S-EPMC6736430 | biostudies-literature
Project description:BackgroundThe usefulness of syndromic surveillance for early outbreak detection depends in part on effective statistical aberration detection. However, few published studies have compared different detection algorithms on identical data. In the largest simulation study conducted to date, we compared the performance of six aberration detection algorithms on simulated outbreaks superimposed on authentic syndromic surveillance data.MethodsWe compared three control-chart-based statistics, two exponential weighted moving averages, and a generalized linear model. We simulated 310 unique outbreak signals, and added these to actual daily counts of four syndromes monitored by Public Health--Seattle and King County's syndromic surveillance system. We compared the sensitivity of the six algorithms at detecting these simulated outbreaks at a fixed alert rate of 0.01.ResultsStratified by baseline or by outbreak distribution, duration, or size, the generalized linear model was more sensitive than the other algorithms and detected 54% (95% CI = 52%-56%) of the simulated epidemics when run at an alert rate of 0.01. However, all of the algorithms had poor sensitivity, particularly for outbreaks that did not begin with a surge of cases.ConclusionWhen tested on county-level data aggregated across age groups, these algorithms often did not perform well in detecting signals other than large, rapid increases in case counts relative to baseline levels.
| S-EPMC1821319 | biostudies-literature
Project description:BackgroundSolomon Islands is one of the least developed countries in the world. Recognising that timely detection of outbreaks is needed to enable early and effective response to disease outbreaks, the Solomon Islands government introduced a simple syndromic surveillance system in 2011. We conducted the first evaluation of the system and the first exploration of a national experience within the broader multi-country Pacific Syndromic Surveillance System to determine if it is meeting its objectives and to identify opportunities for improvement.MethodsWe used a multi-method approach involving retrospective data collection and statistical analysis, modelling, qualitative research and observational methods.ResultsWe found that the system was well accepted, highly relied upon and designed to account for contextual limitations. We found the syndromic algorithm used to identify outbreaks was moderately sensitive, detecting 11.8% (IQR: 6.3-25.0%), 21.3% (IQR: 10.3-36.8%), 27.5% (IQR: 12.8-52.3%) and 40.5% (IQR: 13.5-65.7%) of outbreaks that caused small, moderate, large and very large increases in case presentations to health facilities, respectively. The false alert rate was 10.8% (IQR: 4.8-24.5%). Rural coverage of the system was poor. Limited workforce, surveillance resourcing and other 'upstream' health system factors constrained performance.ConclusionsThe system has made a significant contribution to public health security in Solomon Islands, but remains insufficiently sensitive to detect small-moderate sized outbreaks and hence should not be relied upon as a stand-alone surveillance strategy. Rather, the system should sit within a complementary suite of early warning surveillance activities including event-based, in-patient- and laboratory-based surveillance methods. Future investments need to find a balance between actions to address the technical and systems issues that constrain performance while maintaining simplicity and hence sustainability.
| S-EPMC6302379 | biostudies-literature
Project description:Between August-2016 and April-2017, Solomon Islands experienced the largest and longest-running dengue outbreak on record in the country, with 12,329 suspected cases, 877 hospitalisations and 16 deaths. We conducted a retrospective review of related data and documents, and conducted key informant interviews to characterise the event and investigate the adaptability of syndromic surveillance for enhanced and expanded data collection during a public health emergency in a low resource country setting. While the outbreak quickly consumed available public and clinical resources, we found that authorities were able to scale up the conventional national syndrome-based early warning surveillance system to support the increased information demands during the event demonstrating the flexibility of the system and syndromic surveillance more broadly. Challenges in scaling up included upskilling and assisting staff with no previous experience of the tasks required; managing large volumes of data; maintaining data quality for the duration of the outbreak; harmonising routine and enhanced surveillance data and maintaining surveillance for other diseases; producing information optimally useful for response planning; and managing staff fatigue. Solomon Islands, along with other countries of the region remains vulnerable to outbreaks of dengue and other communicable diseases. Ensuring surveillance systems are robust and able to adapt to changing demands during emergencies should be a health protection priority.
| S-EPMC5991673 | biostudies-literature
Project description:BackgroundThe achievement of the World Health Organization's END TB goals will depend on the successful implementation of strategies for early diagnosis and retention of patients on effective therapy until cure. An estimated 150,000 cases are missed annually in South Africa. It is necessary to look at means for identifying these missed cases. This requires the implementation of active surveillance for TB, a policy adopted by the National Department of Health.AimTo explore the views of managers of the TB program on the implementation of active surveillance for TB in the resource constrained setting of the Eastern Cape, South Africa.MethodsA descriptive, explorative, thematically analysed qualitative study based on 10 semi-structured interviews of managers of the TB program. Interviews were transcribed verbatim and analysed using the framework method and Atlas-ti.ResultsActive case finding of people attending health facilities was the dominant approach, although screening by community health workers (CHWs) was available. Both government and non-government organisations employed CHWs to screen door to door and sometimes as part of campaigns or community events. Some CHWs focused only on contact tracing or people that were non-adherent to TB treatment. Challenges for CHWs included poor coordination and duplication of services, failure to investigate those identified in the community, lack of transport and supportive supervision as well as security issues. Successes included expanding coverage by government CHW teams, innovations to improve screening, strategies to improve CHW capability and attention to social determinants.ConclusionA multifaceted facility- and community-based approach was seen as ideal for active surveillance. More resources should be targeted at strengthening teams of CHWs, for whom this would be part of a comprehensive and integrated service in a community-orientated primary care framework, and community engagement to strengthen community level interventions.
| S-EPMC7531829 | biostudies-literature
Project description:PurposeMisinformation and lack of information about cancer and its treatment pose significant challenges to delivering cancer care in resource-limited settings and may undermine patient engagement in care. We aimed to investigate patients' knowledge and attitudes toward cancer and its treatment and to adapt, implement, and evaluate a low-literacy cancer patient education booklet at the Hôpital Universitaire de Mirebalais (HUM) in rural Haiti.Materials and methodsA low-literacy cancer patient education booklet was adapted into Haitian Creole in collaboration with clinicians at HUM. Patients were recruited for structured interviews (n = 20) and two focus groups (n = 13) designed to explore patients' attitudes toward cancer and its treatment and to assess whether the booklet increased patients' knowledge via an investigator-designed knowledge test.ResultsParticipants reported a subjective lack of knowledge about cancer and its treatments and described views of cancer as deadly or incurable. Patients of varying education levels valued receiving written materials that set expectations about cancer treatment and expressed a desire to share the booklet with caregivers and others in their community. Participants across all levels of education significantly increased their performance on a knowledge test after counseling using the booklet (p < .001).ConclusionWe found that an educational booklet about cancer developed in collaboration with local providers was well received by patients with variable literacy levels and improved their knowledge of cancer and its treatment in a resource-limited setting. Such educational materials have the potential to serve as tools to engage patients with cancer and their families in care.Implications for practiceMisinformation and lack of information pose significant challenges to delivering cancer care in resource-limited settings; however, there are often no culturally and literacy appropriate tools available to aid in patient education. This article shows that written educational materials are well received by patients of variable literacy levels and can be effective tools for increasing patients' knowledge of cancer and its treatment in a limited-resource setting. Furthermore, the authors have made their educational booklet, Cancer and You, freely available online and welcome the opportunity to connect with readers of The Oncologist interested in implementing this educational booklet in clinical care.
| S-EPMC7938404 | biostudies-literature
Project description:School-based influenza-like-illness (ILI) syndromic surveillance can be an important part of influenza community surveillance by providing early warnings for outbreaks and leading to a fast response. From September 2012 to December 2014, syndromic surveillance of ILI was carried out in 4 county-level schools. The cumulative sum methods(CUSUM) was used to detect abnormal signals. A susceptible-exposed-infectious/asymptomatic-recovered (SEIAR) model was fit to the influenza outbreak without control measures and compared with the actual influenza outbreak to evaluate the effectiveness of early control efforts. The ILI incidence rates in 2014 (14.51%) was higher than the incidence in 2013 (5.27%) and 2012 (3.59%). Ten school influenza outbreaks were detected by CUSUM. Each outbreak had high transmissibility with a median Runc of 4.62. The interventions in each outbreak had high effectiveness and all Rcon were 0. The early intervention had high effectiveness within the school-based ILI syndromic surveillance. Syndromic surveillance within schools can play an important role in controlling influenza outbreaks.
| S-EPMC5590954 | biostudies-literature
Project description:BackgroundAs Mozambique faces a double burden of diseases, with a rise of Non Communicable Diseases (NCD) superimposed to uncontrolled communicable diseases (CD), routine disease surveillance system does not include NCD. The objectives of our study were to i) upgrade of the current surveillance system by adapting the data collection tools to NCD; ii) describe the occurrence and profile of selected NCD using these data collection tools.MethodsWorkshops were implemented in a first referral urban hospital of Mozambique to train clinical staff, administrative workers and nurses on NCD surveillance, as well as select conditions to be prioritized. Based on the WHO Global Action Plan and Brazaville Declaration for NCD prevention and control, we selected arterial hypertension, diabetes, stroke, chronic respiratory diseases, mental illness and cancers. Data collection tools used for CD were changed to include age, gender, outcome and visit type. Between February/2014 and January/2015 we collected data at an urban hospital in Mozambique's capital.ResultsOver 12 months 92,018 new patients were assisted in this hospital. Data was missing or diagnosis was unreadable in 2637 (2.9%) thus only 89,381 were used for analysis; of these 6423 (median age 27 years; 58.4% female) had at least one selected NCD as their primary diagnosis: arterial hypertension (2397;37.31%), mental illness (1497;23.30%), asthma (1495;23.28%), diabetes (628;9.78%), stroke (299;4.66%), chronic obstructive pulmonary disease 61 (0.95%) and cancers 46 (0.72%). Emergency transfers were needed for 76 patients (1.2%), mainly due to hypertensive emergencies (31; 40.8%) and stroke (18;23.7%). Twenty-four patients died at entry points (0.3%); 10 of them had hypertensive emergencies.ConclusionChanges in existing surveillance tools for communicable diseases provided important data on the burden and outcomes of the selected NCD helping to identify priority areas for training and health care improvement. This information can be used to design the local NCD clinics and to strengthen the health information system in resource-limited settings in a progressive and sustainable way.
| S-EPMC6360799 | biostudies-literature
Project description:The objective of this study was to assess the performance of several algorithms for outbreak detection based on weekly proportions of whole carcass condemnations. Data from one French slaughterhouse over the 2005-2009 period were used (177 098 slaughtered cattle, 0.97% of whole carcass condemnations). The method involved three steps: (i) preparation of an outbreak-free historical baseline over 5 years, (ii) simulation of over 100 years of baseline time series with injection of artificial outbreak signals with several shapes, durations and magnitudes, and (iii) assessment of the performance (sensitivity, specificity, outbreak detection precocity) of several algorithms to detect these artificial outbreak signals. The algorithms tested included the Shewart p chart, confidence interval of the negative binomial model, the exponentially weighted moving average (EWMA); and cumulative sum (CUSUM). The highest sensitivity was obtained using a negative binomial algorithm and the highest specificity with CUSUM or EWMA. EWMA sensitivity was too low to select this algorithm for efficient outbreak detection. CUSUM's performance was complementary to the negative binomial algorithm. The use of both algorithms on real data for a prospective investigation of the whole carcass condemnation rate as a syndromic surveillance indicator could be relevant. Shewart could also be a good option considering its high sensitivity and simplicity of implementation.
| S-EPMC9151057 | biostudies-literature
Project description:Effective public health surveillance requires consistent monitoring of disease signals such that researchers and decision-makers can react dynamically to changes in disease occurrence. However, whilst surveillance initiatives exist in production animal veterinary medicine, comparable frameworks for companion animals are lacking. First-opinion veterinary electronic health records (EHRs) have the potential to reveal disease signals and often represent the initial reporting of clinical syndromes in animals presenting for medical attention, highlighting their possible significance in early disease detection. Yet despite their availability, there are limitations surrounding their free text-based nature, inhibiting the ability for national-level mortality and morbidity statistics to occur. This paper presents PetBERT, a large language model trained on over 500 million words from 5.1 million EHRs across the UK. PetBERT-ICD is the additional training of PetBERT as a multi-label classifier for the automated coding of veterinary clinical EHRs with the International Classification of Disease 11 framework, achieving F1 scores exceeding 83% across 20 disease codings with minimal annotations. PetBERT-ICD effectively identifies disease outbreaks, outperforming current clinician-assigned point-of-care labelling strategies up to 3 weeks earlier. The potential for PetBERT-ICD to enhance disease surveillance in veterinary medicine represents a promising avenue for advancing animal health and improving public health outcomes.
| S-EPMC10590382 | biostudies-literature