Dealing with uncertainty when using a surveillance system

Using MALDI-TOF spectra in epidemiological surveillance for the detection of bacterial subgroups with a possible epidemic potential

Background

For the purpose of epidemiological surveillance, the Hospital University Institute Méditerranée infection has implemented since 2013 a system named MIDaS, based on the systematic collection of routine activity materials, including MALDI-TOF spectra, and results. The objective of this paper is to present the pipeline we use for processing MALDI-TOF spectra during epidemiological surveillance in order to disclose proteinic cues that may suggest the existence of epidemic processes in complement of incidence surveillance. It is illustrated by the analysis of an alarm observed for Streptococcus pneumoniae.

Methods

The MALDI-TOF spectra analysis process looks for the existence of clusters of spectra characterized by a double time and proteinic close proximity. This process relies on several specific methods aiming at contrasting and clustering the spectra, presenting graphically the results for an easy epidemiological interpretation, and for determining the discriminating spectra peaks with their possible identification using reference databases.

Results

The use of this pipeline in the case of an alarm issued for Streptococcus pneumoniae has made it possible to reveal a cluster of spectra with close proteinic and temporal distances, characterized by the presence of three discriminant peaks (5228.8, 5917.8, and 8974.3 m/z) and the absence of peak 4996.9 m/z. A further investigation on UniProt KB showed that peak 5228.8 is possibly an OxaA protein and that the absent peak may be a transposase.

Conclusion

This example shows this pipeline may support a quasi-real time identification and characterization of clusters that provide essential information on a potentially epidemic situation. It brings valuable information for epidemiological sensemaking and for deciding on the continuation of the epidemiological investigation, in particular the involving of additional costly resources to confirm or invalidate the alarm.

Clinical trials registration

NCT03626987.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12879-021-06803-3.

Estimating COVID-19 cases and outbreaks on-stream through phone calls

One of the main problems in controlling COVID-19 epidemic spread is the delay in confirming cases. Having information on changes in the epidemic evolution or outbreaks rise before laboratory-confirmation is crucial in decision making for Public Health policies. We present an algorithm to estimate on-stream the number of COVID-19 cases using the data from telephone calls to a COVID-line. By modelling the calls as background (proportional to population) plus signal (proportional to infected), we fit the calls in Province of Buenos Aires (Argentina) with coefficient of determination R ² > 0.85. This result allows us to estimate the number of cases given the number of calls from a specific district, days before the laboratory results are available. We validate the algorithm with real data. We show how to use the algorithm to track on-stream the epidemic, and present the Early Outbreak Alarm to detect outbreaks in advance of laboratory results. One key point in the developed algorithm is a detailed track of the uncertainties in the estimations, since the alarm uses the significance of the observables as a main indicator to detect an anomaly. We present the details of the explicit example in Villa Azul (Quilmes) where this tool resulted crucial to control an outbreak on time. The presented tools have been designed in urgency with the available data at the time of the development, and therefore have their limitations which we describe and discuss. We consider possible improvements on the tools, many of which are currently under development.

Decision-making during nonroutine outbreak management: Toward an exploration of experts' creative decisions

Decision-making during critical outbreak management may require standard strategies, but also more creative ones. Our goal was to characterize the expert decision processes that take place during critical situations, where rule-based strategies and usual procedures cannot be satisfactorily applied. More specifically, we focused on the strategies experts use to deal with epidemiological problems, depending on the complexity of the situation. To this end, we carried out a simulated outbreak alert, to place two experts in a situation of epidemiological problem management, based on usual practice but also conducive to implementing creative solutions. To analyze the data, we considered not only the relevance of the solutions proposed by the experts, but also the four creativity criteria defined by Torrance (fluency, flexibility, elaboration and originality). Results allowed us to identify similarities but also differences between the solutions proposed by the experts, depending on their level of experience in this area.

Using decision fusion methods to improve outbreak detection in disease surveillance

BACKGROUND

When outbreak detection algorithms (ODAs) are considered individually, the task of outbreak detection can be seen as a classification problem and the ODA as a sensor providing a binary decision (outbreak yes or no) for each day of surveillance. When they are considered jointly (in cases where several ODAs analyze the same surveillance signal), the outbreak detection problem should be treated as a decision fusion (DF) problem of multiple sensors.

METHODS

This study evaluated the benefit for a decisions support system of using DF methods (fusing multiple ODA decisions) compared to using a single method of outbreak detection. For each day, we merged the decisions of six ODAs using 5 DF methods (two voting methods, logistic regression, CART and Bayesian network - BN). Classical metrics of accuracy, prediction and timelines were used during the evaluation steps.

RESULTS

In our results, we observed the greatest gain (77%) in positive predictive value compared to the best ODA if we used DF methods with a learning step (BN, logistic regression, and CART).

CONCLUSIONS

To identify disease outbreaks in systems using several ODAs to analyze surveillance data, we recommend using a DF method based on a Bayesian network. This method is at least equivalent to the best of the algorithms considered, regardless of the situation faced by the system. For those less familiar with this kind of technique, we propose that logistic regression be used when a training dataset is available.