Timeliness Of Detection Research Articles

Exploring the Benefits of Molecular Testing for Gonorrhoea Antibiotic Resistance Surveillance in Remote Settings.

BackgroundSurveillance for gonorrhoea antimicrobial resistance (AMR) is compromised by a move away from culture-based testing in favour of more convenient nucleic acid amplification test (NAAT) tests. We assessed the potential benefit of a molecular resistance test in terms of the timeliness of detection of gonorrhoea AMR.Methods and FindingsAn individual-based mathematical model was developed to describe the transmission of gonorrhoea in a remote Indigenous population in Australia. We estimated the impact of the molecular test on the time delay between first importation and the first confirmation that the prevalence of gonorrhoea AMR (resistance proportion) has breached the WHO-recommended 5% threshold (when a change in antibiotic should occur). In the remote setting evaluated in this study, the model predicts that when culture is the only available means of testing for AMR, the breach will only be detected when the actual prevalence of AMR in the population has already reached 8 – 18%, with an associated delay of ~43 – 69 months between first importation and detection. With the addition of a molecular resistance test, the number of samples for which AMR can be determined increases facilitating earlier detection at a lower resistance proportion. For the best case scenario, where AMR can be determined for all diagnostic samples, the alert would be triggered at least 8 months earlier than using culture alone and the resistance proportion will have only slightly exceeded the 5% notification threshold.ConclusionsMolecular tests have the potential to provide more timely warning of the emergence of gonorrhoea AMR. This in turn will facilitate earlier treatment switching and more targeted treatment, which has the potential to reduce the population impact of gonorrhoea AMR.

Control over dynamics of large forest fires and assessment of efficiency and timeliness of their detection and suppression

The paper considers methods for assessing the timeliness of detection and the efficiency of the suppression of large forest fires based on a complex analysis of their dynamics, meteorology data, and involved resources. The methods were implemented via the ArcGIS GIS application, which enables the dynamics of each fire to be analyzed as it develops and reports the results of the analysis in the table, graphic (diagram), and cartographic formats.

Hand, foot and mouth disease in China: evaluating an automated system for the detection of outbreaks

ObjectiveTo evaluate the performance of China’s infectious disease automated alert and response system in the detection of outbreaks of hand, foot and mouth (HFM) disease.MethodsWe estimated size, duration and delay in reporting HFM disease outbreaks from cases notified between 1 May 2008 and 30 April 2010 and between 1 May 2010 and 30 April 2012, before and after automatic alert and response included HFM disease. Sensitivity, specificity and timeliness of detection of aberrations in the incidence of HFM disease outbreaks were estimated by comparing automated detections to observations of public health staff.FindingsThe alert and response system recorded 106 005 aberrations in the incidence of HFM disease between 1 May 2010 and 30 April 2012 – a mean of 5.6 aberrations per 100 days in each county that reported HFM disease. The response system had a sensitivity of 92.7% and a specificity of 95.0%. The mean delay between the reporting of the first case of an outbreak and detection of that outbreak by the response system was 2.1 days. Between the first and second study periods, the mean size of an HFM disease outbreak decreased from 19.4 to 15.8 cases and the mean interval between the onset and initial reporting of such an outbreak to the public health emergency reporting system decreased from 10.0 to 9.1 days.ConclusionThe automated alert and response system shows good sensitivity in the detection of HFM disease outbreaks and appears to be relatively rapid. Continued use of this system should allow more effective prevention and limitation of such outbreaks in China.

Modelling Reporting Delays for Outbreak Detection in Infectious Disease Data

SummaryThe delay that necessarily occurs between the emergence of symptoms and the identification of the cause of those symptoms affects the timeliness of detection of emerging outbreaks of infectious diseases, and hence the ability to take preventive action. We study the delays that are associated with the collection of laboratory surveillance data in England, Wales and Northern Ireland, using 12 infections of contrasting characteristics. We use a continuous time spline-based model for the hazard of the delay distribution, along with an associated proportional hazards model. The delay distributions are found to have extremely long tails, the hazard at longer delays being roughly constant, suggestive of a memoryless process, though some laboratories appear to stop reporting after a certain delay. The hazards are found typically to vary strongly with calendar time, and to a lesser extent with season and recent organism frequency. In consequence, the delay distributions cannot be assumed to be stationary. These findings will inform the development of outbreak detection algorithms that take account of reporting delays.

Use of monthly collected milk yields for the detection of the emergence of the 2007 French BTV epizootic

Two culicoides-borne diseases, Bluetongue (BTV) and Schmallenberg, have emerged in the European cattle population since 2006. Other diseases transmitted by these vectors could emerge. This justifies the development of syndromic surveillance programs whereby one or several indicators would be routinely monitored for the early detection of emerging diseases. The aim of this study was to evaluate milk yield from milk recording in dairy cattle as an indicator to be included in an emerging disease surveillance system. It was hypothesized that emergences would result in episodes of low milk production clustered in space and time. The 2007 BTV epizootic in France was used as a case study. Because it had already emerged in neighbouring countries, the disease emergence was expected and notification was mandatory. Herd-test-day milk productions were predicted for the entire country for 2006 and 2007 from herd historical data using linear mixed models. The differences between observed and predicted milk productions were averaged per week and per municipality and used as input for a space-time prospective scan statistic. Log likelihood ratios (LLR) associated with clusters were used to define alarms. The threshold chosen was a trade-off between detection timeliness and the number of false alarms per week. The first four BTV notifications occurred on the 12th (two notifications), 13th and 27th of July 2007. The 12th of July was considered to be the date of emergence. Alarms occurring before the 1st of March 2007 were considered to be false alarms. Using an LLR of 50, there were an average of 1.7 false alarms per week and the BTV emergence was detected seven weeks after emergence. Using an LLR of 100, there were an average of 0.8 false alarms per week and the BTV emergence was detected 9 weeks after emergence. Detection may have been delayed because of a discontinuation of milk recording between mid-July and mid-August. The first cluster with an LLR>100 located in the emergence area was further investigated. A difference between observed and predicted production of >1kg/cow/day was observed around the time of emergence. However, a difference of equal magnitude was observed during the year preceding the outbreak. Milk production predicted from herd history alone did not allow the detection of the 2007 BTV emergence in France. Further research should be conducted on improving the prediction of test-day milk yield and on combining it with other indicators based on routinely collected data.

Evaluation of a Continuous Indicator for Syndromic Surveillance through Simulation. Application to Vector Borne Disease Emergence Detection in Cattle Using Milk Yield

Two vector borne diseases, caused by the Bluetongue and Schmallenberg viruses respectively, have emerged in the European ruminant populations since 2006. Several diseases are transmitted by the same vectors and could emerge in the future. Syndromic surveillance, which consists in the routine monitoring of indicators for the detection of adverse health events, may allow an early detection. Milk yield is routinely measured in a large proportion of dairy herds and could be incorporated as an indicator in a surveillance system. However, few studies have evaluated continuous indicators for syndromic surveillance. The aim of this study was to develop a framework for the quantification of both disease characteristics and model predictive abilities that are important for a continuous indicator to be sensitive, timely and specific for the detection of a vector-borne disease emergence. Emergences with a range of spread characteristics and effects on milk production were simulated. Milk yields collected monthly in 48 713 French dairy herds were used to simulate 576 disease emergence scenarios. First, the effect of disease characteristics on the sensitivity and timeliness of detection were assessed: Spatio-temporal clusters of low milk production were detected with a scan statistic using the difference between observed and simulated milk yields as input. In a second step, the system specificity was evaluated by running the scan statistic on the difference between observed and predicted milk yields, in the absence of simulated emergence. The timeliness of detection depended mostly on how easily the disease spread between and within herds. The time and location of the emergence or adding random noise to the simulated effects had a limited impact on the timeliness of detection. The main limitation of the system was the low specificity i.e. the high number of clusters detected from the difference between observed and predicted productions, in the absence of disease.

Application of the China Infectious Diseases Automated-alert and Response System in Guangxi, 2009-2011

To analyze and further improvement the application of the China Infectious Diseases Automated-alert and Response System (CIDARS) in Guangxi Zhuang Autonomous Region. Results related to the amount of signal, proportion of signal responded, time to signal response, manner of signal verification and on each signal of Guangxi in CIDARS from 2009 to 2011 were described. Performance was compared between the periods of pre/ post the adjustment of parameters in CIDARS on December 10, 2010. A total of 29 788 signals were generated on 16 infectious diseases in the system in Guangxi. 100% signals had been responded with the median time to response as 1.5 hours. The average amount of signal per county per week was 1.7;with 624 signals(2.09%)verified as suspected outbreaks preliminarily and 191 outbreaks of 9 diseases were finally confirmed by further field investigation. The sensitivity of CIDARS was 89.25% , and the timeliness of detection was 2.8 d. After adjusting the parameter of CIDARS, the number of signals reduced, and the sensitivity and timeliness of detection improved for most of the diseases. The signals of CIDARS were responded timely, and the performance of CIDARS might be improved by adjusting the parameters of early-warning model, which helped enhance the ability of outbreaks-detection for local public health departments. However the current proportion of false positive signals still seemed to be high, suggesting that both the methods and parameters should be improved, according to the characteristics of different diseases.

Impact of United States Food and Drug Administration's boxed warnings on adverse drug reactions reporting rates and risk mitigation for multiple myeloma drugs

Purpose:To determine the relationship between boxed warnings issuance by the US Food and Drug Administration (FDA) and the proportional reporting rates of the associated adverse drug reactions (ADRs) to the FDA's Adverse Event Reporting System (FAERS) for multiple myeloma (MM) drugs.Methods:We compiled a list of all FDA approved MM drugs and identified their associated ADR boxed warnings, through FDA's website and physician desk reference. Drugs that were issued boxed warnings after their market launch were included in the analysis, i.e., melphalan, thalidomide, vincristine, carmustine and doxorubicin. For each drug/ADR boxed warning combination, we retrieved all reported cases from the FAERS and calculated their Empiric Bayes Geometric Means (EBGMs), in pre- and post-boxed warning periods. Chi-square tests were performed to compare serious adverse drug events before and after boxed warnings for all drug/ADR combinations.Results:A total of 10 drug/ADR boxed warning combinations were identified, of which EBGM signals increased for six combinations after a boxed warning was issued. Reports of serious adverse drug events also increased significantly (p < 0.05).Conclusion:Boxed warnings were associated with increased FAERS reporting, indicating increased awareness of ADRs for MM drugs. Proactive pharmacovigilance programs, such as the FDA's Mini-Sentinel Project, may improve timeliness of detection of rare ADRs.

In silico Surveillance: evaluating outbreak detection with simulation models

BackgroundDetecting outbreaks is a crucial task for public health officials, yet gaps remain in the systematic evaluation of outbreak detection protocols. The authors’ objectives were to design, implement, and test a flexible methodology for generating detailed synthetic surveillance data that provides realistic geographical and temporal clustering of cases and use to evaluate outbreak detection protocols.MethodsA detailed representation of the Boston area was constructed, based on data about individuals, locations, and activity patterns. Influenza-like illness (ILI) transmission was simulated, producing 100 years of in silico ILI data. Six different surveillance systems were designed and developed using gathered cases from the simulated disease data. Performance was measured by inserting test outbreaks into the surveillance streams and analyzing the likelihood and timeliness of detection.ResultsDetection of outbreaks varied from 21% to 95%. Increased coverage did not linearly improve detection probability for all surveillance systems. Relaxing the decision threshold for signaling outbreaks greatly increased false-positives, improved outbreak detection slightly, and led to earlier outbreak detection.ConclusionsGeographical distribution can be more important than coverage level. Detailed simulations of infectious disease transmission can be configured to represent nearly any conceivable scenario. They are a powerful tool for evaluating the performance of surveillance systems and methods used for outbreak detection.

Conjugate vaccine introduction in the African meningitis belt: meeting surveillance objectives

The epidemiology of bacterial meningitis will change with introduction of meningococcal and pneumococcal conjugate vaccines in the African meningitis belt. The principal objectives of surveillance are to evaluate the impact of vaccination, to detect and investigate epidemics and provide material for research. The capacity of existing surveillance activities in the meningitis belt to meet these objectives varies due to infrastructural and financial constraints. Impact assessment of conjugate vaccine against meningococcal serogroup A will be limited to comparing incidence trends from a few surveillance sites with data obtained before vaccine introduction and to comparing trends in the incidence of suspected cases or localised epidemics in most other settings. The timeliness of detection of epidemics and identification of epidemic meningococcal serogroups could be improved in most countries by analysing suspected case data in health centre level resolution and by investigating outbreaks with mobile teams. For research and impact assessment of pneumococcal conjugate vaccines, several surveillance sites covering at least 0.5 million inhabitants should be maintained which undertake exhaustive case finding and systematic laboratory confirmation of meningitis and pneumonia. Molecular diagnostics will facilitate surveillance in remote areas, but the available techniques should be evaluated for diagnostic performance in the field and long-term sustainability.

On timeliness and accuracy of wildfire detection by the GOES WF-ABBA algorithm over California during the 2006 fire season

The Wildfire Automated Biomass Burning Algorithm (WF-ABBA) is a state-of-the-art algorithm for geostationary wildfire detection whose results have been increasingly used in a range of environmental applications. At present, the WF-ABBA validation activities and, in general, fire product validation methodologies are at a markedly less advanced stage than the algorithm itself. Particularly, little is known about detection timeliness, despite the value of such information for assessing the potential of geostationary observations to improve tactical decision making of first responders. This paper contributes to reducing this gap in two ways. Firstly, we describe a new methodology that is suitable for evaluating geostationary satellite wildfire detection in terms of incidents with regard to both timeliness and reliability. This methodology utilizes available official multi-agency wildfire reporting information and multitemporal Landsat imagery. Secondly, we apply the proposed validation method to temporally filtered GOES-West WF-ABBA (ver. 6.1) detections for the 2006 fire season over the State of California and present incident-wise and pixel-wise performance information. The results indicate highly reliable pixel-wise performance of WF-ABBA, with about 75% of fire pixels (or more) corresponding to actual recorded active wildfires. A substantial portion of wildfires were detected during their first hour of activity, and a few incidents—even before the initial reports from conventional sources. Although the WF-ABBA performs best at what it was designed for: consistently re-detecting (monitoring) active fires, we believe there is an additional potential for automated detection from current geostationary data to reduce wildfire ignition latencies in the Western U.S. Our results can serve as a guideline for algorithm developers and users of the WF-ABBA fire product.

Effects of malaria volunteer training on coverage and timeliness of diagnosis: a cluster randomized controlled trial in Myanmar

BackgroundThe use of community volunteers is expected to improve access to accurate diagnosis and timely treatment of malaria, using rapid diagnostic test (RDT) and artemisinin-based combination therapy (ACT). However, empirical data from the field are still limited. The aim of this study was to assess whether training village volunteers on the use of Paracheck-Pf® RDT and ACT (artemether-lumefantrine (AL)) for Plasmodium falciparum and presumptive treatment with chloroquine for Plasmodium vivax had an effect on the coverage of timely diagnosis and treatment and on mortality in malaria-endemic villages without health staff in Myanmar.MethodsThe study was designed as a cluster randomized controlled trial with a cross-sectional survey at baseline, a monthly visit for six months following the intervention (village volunteers trained and equipped with Paracheck-Pf®) and an endline survey at six months follow-up. Survey data were supplemented by the analysis of logbooks and field-based verbal autopsies. Villages with midwives (MW) in post were used as a third comparison group in the endline survey. Intention-to-treat analysis was used.ResultsOf 38 villages selected, 21 were randomly assigned to the intervention (two villages failed to participate) and 17 to the comparison group. The two groups had comparable baseline statistics. The blood tests provided by volunteers every month declined over time from 279 tests to 41 but not in MW group in 18 villages (from 326 to 180). In the endline survey, among interviewed subjects (268 intervention, 287 in comparison, 313 in MW), the coverage of RDT was low in all groups (14.9%, SE 2.4% in intervention; 5.7%, SE 1.7% in comparison; 21.4%, SE 2.6% in MW) although the intervention (OR 3.2, 95% CI 1.5-6.7) and MW (OR 5.4, 95% CI 2.6-11.0) were more likely to receive a blood test. Mean (SE) of blood tests after onset of fever in days was delayed (intervention 3.6 (0.3); comparison 4.8 (1.3); MW 3.2 (0.4)). Malaria mortality rates per 100,000 populations in a year were not significantly different (intervention 130 SE 37; comparison 119 SE 34; MW 50 SE 18). None of the dead cases had consulted volunteers.ConclusionsThe results show that implementing volunteer programmes to improve the coverage of accurate and timely diagnosis with RDT and early treatment may be beneficial but the timeliness of detection and sustainability must be improved.

Evaluating Surveillance Strategies for the Early Detection of Low Pathogenicity Avian Influenza Infections

In recent years, the early detection of low pathogenicity avian influenza (LPAI) viruses in poultry has become increasingly important, given their potential to mutate into highly pathogenic viruses. However, evaluations of LPAI surveillance have mainly focused on prevalence and not on the ability to act as an early warning system. We used a simulation model based on data from Italian LPAI epidemics in turkeys to evaluate different surveillance strategies in terms of their performance as early warning systems. The strategies differed in terms of sample size, sampling frequency, diagnostic tests, and whether or not active surveillance (i.e., routine laboratory testing of farms) was performed, and were also tested under different epidemiological scenarios. We compared surveillance strategies by simulating within-farm outbreaks. The output measures were the proportion of infected farms that are detected and the farm reproduction number (Rh). The first one provides an indication of the sensitivity of the surveillance system to detect within-farm infections, whereas Rh reflects the effectiveness of outbreak detection (i.e., if detection occurs soon enough to bring an epidemic under control). Increasing the sampling frequency was the most effective means of improving the timeliness of detection (i.e., it occurs earlier), whereas increasing the sample size increased the likelihood of detection. Surveillance was only effective in preventing an epidemic if actions were taken within two days of sampling. The strategies were not affected by the quality of the diagnostic test, although performing both serological and virological assays increased the sensitivity of active surveillance. Early detection of LPAI outbreaks in turkeys can be achieved by increasing the sampling frequency for active surveillance, though very frequent sampling may not be sustainable in the long term. We suggest that, when no LPAI virus is circulating yet and there is a low risk of virus introduction, a less frequent sampling approach might be admitted, provided that the surveillance is intensified as soon as the first outbreak is detected.

Rank-based spatial clustering: an algorithm for rapid outbreak detection

Public health surveillance requires outbreak detection algorithms with computational efficiency sufficient to handle the increasing volume of disease surveillance data. In response to this need, the authors propose a spatial clustering algorithm, rank-based spatial clustering (RSC), that detects rapidly infectious but non-contagious disease outbreaks. The authors compared the outbreak-detection performance of RSC with that of three well established algorithms-the wavelet anomaly detector (WAD), the spatial scan statistic (KSS), and the Bayesian spatial scan statistic (BSS)-using real disease surveillance data on to which they superimposed simulated disease outbreaks. The following outbreak-detection performance metrics were measured: receiver operating characteristic curve, activity monitoring operating curve curve, cluster positive predictive value, cluster sensitivity, and algorithm run time. RSC was computationally efficient. It outperformed the other two spatial algorithms in terms of detection timeliness, and outbreak localization. RSC also had overall better timeliness than the time-series algorithm WAD at low false alarm rates. RSC is an ideal algorithm for analyzing large datasets when the application of other spatial algorithms is not practical. It also allows timely investigation for public health practitioners by providing early detection and well-localized outbreak clusters.

Fast Bayesian scan statistics for multivariate event detection and visualization

The multivariate Bayesian scan statistic (MBSS) is a recently proposed, general framework for event detection and characterization in multivariate space-time data. MBSS integrates prior information and observations from multiple data streams in a Bayesian framework, computing the posterior probability of each type of event in each space-time region. MBSS has been shown to have many advantages over previous event detection approaches, including improved timeliness and accuracy of detection, easy interpretation and visualization of results, and the ability to model and accurately differentiate between multiple event types. This work extends the MBSS framework to enable detection and visualization of irregularly shaped clusters in multivariate data, by defining a hierarchical prior over all subsets of locations. While a naive search over the exponentially many subsets would be computationally infeasible, we demonstrate that the total posterior probability that each location has been affected can be efficiently computed, enabling rapid detection and visualization of irregular clusters. We compare the run time and detection power of this 'Fast Subset Sums' method to our original MBSS approach (assuming a uniform prior over circular regions) on semi-synthetic outbreaks injected into real-world Emergency Department data from Allegheny County, Pennsylvania. We demonstrate substantial improvements in spatial accuracy and timeliness of detection, while maintaining the scalability and fast run time of the original MBSS method.

Use of outcomes to evaluate surveillance systems for bioterrorist attacks

BackgroundSyndromic surveillance systems can potentially be used to detect a bioterrorist attack earlier than traditional surveillance, by virtue of their near real-time analysis of relevant data. Receiver operator characteristic (ROC) curve analysis using the area under the curve (AUC) as a comparison metric has been recommended as a practical evaluation tool for syndromic surveillance systems, yet traditional ROC curves do not account for timeliness of detection or subsequent time-dependent health outcomes.MethodsUsing a decision-analytic approach, we predicted outcomes, measured in lives, quality adjusted life years (QALYs), and costs, for a series of simulated bioterrorist attacks. We then evaluated seven detection algorithms applied to syndromic surveillance data using outcomes-weighted ROC curves compared to simple ROC curves and timeliness-weighted ROC curves. We performed sensitivity analyses by varying the model inputs between best and worst case scenarios and by applying different methods of AUC calculation.ResultsThe decision analytic model results indicate that if a surveillance system was successful in detecting an attack, and measures were immediately taken to deliver treatment to the population, the lives, QALYs and dollars lost could be reduced considerably. The ROC curve analysis shows that the incorporation of outcomes into the evaluation metric has an important effect on the apparent performance of the surveillance systems. The relative order of performance is also heavily dependent on the choice of AUC calculation method.ConclusionsThis study demonstrates the importance of accounting for mortality, morbidity and costs in the evaluation of syndromic surveillance systems. Incorporating these outcomes into the ROC curve analysis allows for more accurate identification of the optimal method for signaling a possible bioterrorist attack. In addition, the parameters used to construct an ROC curve should be given careful consideration.

Study on the timeliness of detection and reporting on public health emergency events in China

To analyze the timeliness of detection and reporting on public health emergency events, and to explore the effective strategies for improving the relative capacity on those issues. We conducted a retrospective survey on 3275 emergency events reported through Public Health Emergency Events Surveillance System from 2005 to the first half of 2006. Developed by county Centers for Disease Control and Prevention, a uniformed self-administrated questionnaire was used to collect data, which would include information on the detection, reporting of the events. For communicable diseases events, the median of time interval between the occurrence of first case and the detection of event was 6 days (P25 = 2, P75 = 13). For food poisoning events and clusters of disease with unknown origin, the medians were 3 hours (P25, P75 = 16) and 1 days (P25 = 0, P75 = 5). 71.54% of the events were reported by the discoverers within 2 hours after the detection. In general, the ranges of time intervals between the occurrence, detection or reporting of the events were different, according to the categories of events. The timeliness of detection and reporting of events could have been improved dramatically if the definition of events, according to their characteristics, had been more reasonable and accessible, as well as the improvement of training program for healthcare staff and teachers.

Early adverse drug event signal detection within population‐based health networks using sequential methods: key methodologic considerations

Active surveillance of population-based health networks may improve the timeliness of detection of adverse events (AEs). Our objective was to expand our previous signal detection work by investigating the effect on signal detection of alternative study specifications. We compared the signal detection performance under various study specifications using historical data from nine health plans involved in the HMO Research Network's Center for Education and Research on Therapeutics (CERT). Five drug-event pairs representing generally accepted associations with an AE and two pairs representing "negative controls" were analyzed. Alternative study specifications related to the definition of incident users and incident AEs were assessed and compared to our previous findings. Relaxing the incident AE exclusion criteria by (1) including members with prior outpatient diagnoses of interest and (2) halving (to 90 days) the time window specified to define incident exposure and diagnoses increased the number of members under surveillance and as a consequence increased the number of exposed days and diagnoses by about 10-20%. The alternative specifications tend to result in earlier signal detection by 10-16 months, a likely consequence of more exposures and events entering the analysis. This paper provides additional preliminary information related to conducting prospective safety monitoring using health plan data and sequential analytic methods. Our findings support continued investigation of using health plan data and sequential analytic methods as a potentially important contribution to active drug safety surveillance.

Longitudinal histories as predictors of future diagnoses of domestic abuse: modelling study

Objective To determine whether longitudinal data in patients’ historical records, commonly available in electronic health record systems, can be used to predict a patient’s future risk of receiving a diagnosis...

Biosurveillance of emerging biothreats using scalable genotype clustering

Developments in molecular fingerprinting of pathogens with epidemic potential have offered new opportunities for improving detection and monitoring of biothreats. However, the lack of scalable definitions for infectious disease clustering presents a barrier for effective use and evaluation of new data types for early warning systems. A novel working definition of an outbreak based on temporal and spatial clustering of molecular genotypes is introduced in this paper. It provides an unambiguous way of clustering of causative pathogens and is adjustable to local disease prevalence and availability of public health resources. The performance of this definition in prospective surveillance is assessed in the context of community outbreaks of food-borne salmonellosis. Molecular fingerprinting augmented with the scalable clustering allows the detection of more than 50% of the potential outbreaks before they reach the midpoint of the cluster duration. Clustering in time by imposing restrictions on intervals between collection dates results in a smaller number of outbreaks but does not significantly affect the timeliness of detection. Clustering in space and time by imposing restrictions on the spatial and temporal distance between cases results in a further reduction in the number of outbreaks and decreases the overall efficiency of prospective detection. Innovative bacterial genotyping technologies can enhance early warning systems for public health by aiding the detection of moderate and small epidemics.