Temporal surveillance using scan statistics

An Emergency Preparedness Response to Opioid-Prescribing Enforcement Actions in Maryland, 2018-2019

Federal and state enforcement authorities have increasingly intervened on the criminal overprescribing of opioids. However, little is known about the health effects these enforcement actions have on patients experiencing disrupted access to prescription opioids or medication-assisted treatment/medication for opioid use disorder. Simultaneously, opioid death rates have increased. In response, the Maryland Department of Health (MDH) has worked to coordinate mitigation strategies with enforcement partners (defined as any federal, state, or local enforcement authority or other governmental investigative authority). One strategy is a standardized protocol to implement emergency response functions, including rapidly identifying health hazards with real-time data access, deploying resources locally, and providing credible messages to partners and the public. From January 2018 through October 2019, MDH used the protocol in response to 12 enforcement actions targeting 34 medical professionals. A total of 9624 patients received Schedule II-V controlled substance prescriptions from affected prescribers under investigation in the 6 months before the respective enforcement action; 9270 (96%) patients were residents of Maryland. Preliminary data indicate fatal overdose events and potential loss of follow-up care among the patient population experiencing disrupted health care as a result of an enforcement action. The success of the strategy hinged on endorsement by leadership; the establishment of federal, state, and local roles and responsibilities; and data sharing. MDH's approach, data sources, and lessons learned may support health departments across the country that are interested in conducting similar activities on the front lines of the opioid crisis.

Identification of a morning out-of-hospital cardiac arrest cluster of high-incidence: towards a chrono-preventive care strategy

RATIONALE, AIMS, AND OBJECTIVES

The human body is regulated by intrinsic factors which follow a 24-hour biological clock. Implications of a circadian rhythm in the out-of-hospital cardiac arrest (OHCA) are studied but the literature is not consistent. The main objective of our study was to identify temporal cluster of high or low incidence of OHCA occurrence during a day.

METHODS

Multicentre comparative study based on the French national OHCA registry data between 2013 and 2017. After describing the population, the detection of significant temporal clusters of OHCA incidence was achieved using temporal scan statistics based on a Poisson model adjusted for age and gender. Then, comparisons between identified patients clusters and the rest of the population were performed.

RESULTS

During the study, 37 163 medical OHCA victims were included. The temporal scan revealed a significant 3-hour high incidence temporal cluster between 8:00 am and 10:59 am (Relative R = 1.76, P < .001). In the identified cluster, OHCA occurred more out of the home with fewer witnesses, and advanced life support was less attempted in the cluster. No difference was observed on the return of spontaneous circulation, survival at hospital admission, and survival 30 days after the OHCA or at hospital discharge.

CONCLUSIONS

We observed a three-hour morning high incidence peak of OHCA. This high incidence could be explained by different physiological changes in the morning. These changes are well known and the evidence of a morning peak of cardiovascular disease should enable medical teams to adapt care strategy and hospital organization.

Enhancing the monitoring of fallen stock at different hierarchical administrative levels: an illustration on dairy cattle from regions with distinct husbandry, demographical and climate traits

Background

The automated collection of non-specific data from livestock, combined with techniques for data mining and time series analyses, facilitates the development of animal health syndromic surveillance (AHSyS). An example of AHSyS approach relates to the monitoring of bovine fallen stock. In order to enhance part of the machinery of a complete syndromic surveillance system, the present work developed a novel approach for modelling in near real time multiple mortality patterns at different hierarchical administrative levels. To illustrate its functionality, this system was applied to mortality data in dairy cattle collected across two Spanish regions with distinct demographical, husbandry, and climate conditions.

Results

The process analyzed the patterns of weekly counts of fallen dairy cattle at different hierarchical administrative levels across two regions between Jan-2006 and Dec-2013 and predicted their respective expected counts between Jan-2014 and Jun- 2015. By comparing predicted to observed data, those counts of fallen dairy cattle that exceeded the upper limits of a conventional 95% predicted interval were identified as mortality peaks. This work proposes a dynamic system that combines hierarchical time series and autoregressive integrated moving average models (ARIMA). These ARIMA models also include trend and seasonality for describing profiles of weekly mortality and detecting aberrations at the region, province, and county levels (spatial aggregations). Software that fitted the model parameters was built using the R statistical packages.

Conclusions

The work builds a novel tool to monitor fallen stock data for different geographical aggregations and can serve as a means of generating early warning signals of a health problem. This approach can be adapted to other types of animal health data that share similar hierarchical structures.

Congenital Anomalies: Cluster Detection and Investigation

This work summarizes the main aspects to be considered around birth defects (or congenital anomalies) clusters. Most birth defects (BD), considered individually, fall into the definition of rare diseases (RD), according to their low frequency. Likewise, many RD are congenital, because their manifestations are present at birth or can be even evident before the delivery. It has been estimated that overall 7.9 million children are born each year with serious BD of genetic or partially genetic origin, and additional hundreds of thousands more are born with serious BD of post-conception origin.A "birth defect cluster" can be defined as an unusual aggregation of cases (grouped in place and time) that is suspected to be greater than expected, even though the expected number may not be known. These clusters are incidents or occurrences that let us turn the challenge of identifying the causal agent(s) involved in the origin of such clusters, into an opportunity to exert primary prevention, and thus achieve the ultimate goal of enabling infants being born healthy. Therefore, any program or system involved in BD surveillance and research should devote part of its activities to detect and investigate clusters, to ensure that such opportunity for primary prevention will be conveniently leveraged. Regardless the type of cluster, there are several phases that must be undertaken sequentially for proper control and the maximum benefit for the population: cluster detection, evaluation and investigation, management, adoption of preventive measures, and communication of the results to the public or target population.

Spatial and Temporal Clustering of Anti-Glomerular Basement Membrane Disease

BACKGROUND AND OBJECTIVES

An environmental trigger has been proposed as an inciting factor in the development of anti-GBM disease. This multicenter, observational study sought to define the national incidence of anti-GBM disease during an 11-year period (2003-2014) in Ireland, investigate clustering of cases in time and space, and assess the effect of spatial variability in incidence on outcome.

DESIGN, SETTING, PARTICIPANTS, & MEASUREMENTS

We ascertained cases by screening immunology laboratories for instances of positivity for anti-GBM antibody and the national renal histopathology registry for biopsy-proven cases. The population at risk was defined from national census data. We used a variable-window scan statistic to detect temporal clustering. A Bayesian spatial model was used to calculate standardized incidence ratios (SIRs) for each of the 26 counties.

RESULTS

Seventy-nine cases were included. National incidence was 1.64 (95% confidence interval [95% CI], 0.82 to 3.35) per million population per year. A temporal cluster (n=10) was identified during a 3-month period; six cases were resident in four rural counties in the southeast. Spatial analysis revealed wide regional variation in SIRs and a cluster (n=7) in the northwest (SIR, 1.71; 95% CI, 1.02 to 3.06). There were 29 deaths and 57 cases of ESRD during a mean follow-up of 2.9 years. Greater distance from diagnosis site to treating center, stratified by median distance traveled, did not significantly affect patient (hazard ratio, 1.80; 95% CI, 0.87 to 3.77) or renal (hazard ratio, 0.76; 95% CI, 0.40 to 1.13) survival.

CONCLUSIONS

To our knowledge, this is the first study to report national incidence rates of anti-GBM disease and formally investigate patterns of incidence. Clustering of cases in time and space supports the hypothesis of an environmental trigger for disease onset. The substantial variability in regional incidence highlights the need for comprehensive country-wide studies to improve our understanding of the etiology of anti-GBM disease.

Detection and investigation of temporal clusters of congenital anomaly in Europe: seven years of experience of the EUROCAT surveillance system

Detection and investigation of congenital anomaly clusters is one part of surveillance to detect new or changing teratogenic exposures in the population. The EUROCAT (European Surveillance of Congenital Anomalies) cluster monitoring system and results are described here. Monitoring was conducted annually from 2007 to 2013 for 18 registries covering an annual birth population up to 0.5 million births. For each registry and 72 anomaly subgroups, the scan "moving window" technique was used to detect clusters in time occurring within the last 2 years based on estimated date of conception. Registries conducted preliminary investigations using a standardised protocol to determine whether there was cause for concern, and expert review was used at key points. 165 clusters were detected, a rate of 3.4% of all 4823 cluster tests performed over 7 years, more than expected by chance. Preliminary investigations of 126 new clusters confirmed that 35% were an unusual aggregation of cases, while 56% were explained by data quality or diagnostic issues, and 9% were not investigated. For confirmed clusters, the registries' course of action was continuing monitoring. Three confirmed clusters continued to grow in size for a limited period in subsequent monitoring. This system is best suited to early detection of exposures which are sudden, widespread and/or highly teratogenic, and was reassuring in demonstrating an absence of a sustained exposure of this type. Such proactive monitoring can be run efficiently without overwhelming the surveillance system with false positives, and serves an additional purpose of data quality control.

Using scan statistics for congenital anomalies surveillance: the EUROCAT methodology

Scan statistics have been used extensively to identify temporal clusters of health events. We describe the temporal cluster detection methodology adopted by the EUROCAT (European Surveillance of Congenital Anomalies) monitoring system. Since 2001, EUROCAT has implemented variable window width scan statistic for detecting unusual temporal aggregations of congenital anomaly cases. The scan windows are based on numbers of cases rather than being defined by time. The methodology is imbedded in the EUROCAT Central Database for annual application to centrally held registry data. The methodology was incrementally adapted to improve the utility and to address statistical issues. Simulation exercises were used to determine the power of the methodology to identify periods of raised risk (of 1-18 months). In order to operationalize the scan methodology, a number of adaptations were needed, including: estimating date of conception as unit of time; deciding the maximum length (in time) and recency of clusters of interest; reporting of multiple and overlapping significant clusters; replacing the Monte Carlo simulation with a lookup table to reduce computation time; and placing a threshold on underlying population change and estimating the false positive rate by simulation. Exploration of power found that raised risk periods lasting 1 month are unlikely to be detected except when the relative risk and case counts are high. The variable window width scan statistic is a useful tool for the surveillance of congenital anomalies. Numerous adaptations have improved the utility of the original methodology in the context of temporal cluster detection in congenital anomalies.

Detection and investigation of temporal clusters of congenital anomaly in Europe: seven years of experience of the EUROCAT surveillance system

Detection and investigation of congenital anomaly clusters is one part of surveillance to detect new or changing teratogenic exposures in the population. The EUROCAT (European Surveillance of Congenital Anomalies) cluster monitoring system and results are described here. Monitoring was conducted annually from 2007 to 2013 for 18 registries covering an annual birth population up to 0.5 million births. For each registry and 72 anomaly subgroups, the scan "moving window" technique was used to detect clusters in time occurring within the last 2 years based on estimated date of conception. Registries conducted preliminary investigations using a standardised protocol to determine whether there was cause for concern, and expert review was used at key points. 165 clusters were detected, a rate of 3.4% of all 4823 cluster tests performed over 7 years, more than expected by chance. Preliminary investigations of 126 new clusters confirmed that 35% were an unusual aggregation of cases, while 56% were explained by data quality or diagnostic issues, and 9% were not investigated. For confirmed clusters, the registries' course of action was continuing monitoring. Three confirmed clusters continued to grow in size for a limited period in subsequent monitoring. This system is best suited to early detection of exposures which are sudden, widespread and/or highly teratogenic, and was reassuring in demonstrating an absence of a sustained exposure of this type. Such proactive monitoring can be run efficiently without overwhelming the surveillance system with false positives, and serves an additional purpose of data quality control.

A critical look at prospective surveillance using a scan statistic

The scan statistic is a very popular surveillance technique for purely spatial, purely temporal, and spatial-temporal disease data. It was extended to the prospective surveillance case, and it has been applied quite extensively in this situation. When the usual signal rules, as those implemented in SaTScan(TM) (Boston, MA, USA) software, are used, we show that the scan statistic method is not appropriate for the prospective case. The reason is that it does not adjust properly for the sequential and repeated tests carried out during the surveillance. We demonstrate that the nominal significance level α is not meaningful and there is no relationship between α and the recurrence interval or the average run length (ARL). In some cases, the ARL may be equal to ∞, which makes the method ineffective. This lack of control of the type-I error probability and of the ARL leads us to strongly oppose the use of the scan statistic with the usual signal rules in the prospective context.

Retrospective time series analysis of veterinary laboratory data: preparing a historical baseline for cluster detection in syndromic surveillance

The practice of disease surveillance has shifted in the last two decades towards the introduction of systems capable of early detection of disease. Modern biosurveillance systems explore different sources of pre-diagnostic data, such as patient's chief complaint upon emergency visit or laboratory test orders. These sources of data can provide more rapid detection than traditional surveillance based on case confirmation, but are less specific, and therefore their use poses challenges related to the presence of background noise and unlabelled temporal aberrations in historical data. The overall goal of this study was to carry out retrospective analysis using three years of laboratory test submissions to the Animal Health Laboratory in the province of Ontario, Canada, in order to prepare the data for use in syndromic surveillance. Daily cases were grouped into syndromes and counts for each syndrome were monitored on a daily basis when medians were higher than one case per day, and weekly otherwise. Poisson regression accounting for day-of-week and month was able to capture the day-of-week effect with minimal influence from temporal aberrations. Applying Poisson regression in an iterative manner, that removed data points above the predicted 95th percentile of daily counts, allowed for the removal of these aberrations in the absence of labelled outbreaks, while maintaining the day-of-week effect that was present in the original data. This resulted in the construction of time series that represent the baseline patterns over the past three years, free of temporal aberrations. The final method was thus able to remove temporal aberrations while keeping the original explainable effects in the data, did not need a training period free of aberrations, had minimal adjustment to the aberrations present in the raw data, and did not require labelled outbreaks. Moreover, it was readily applicable to the weekly data by substituting Poisson regression with moving 95th percentiles.

Bowel sounds and monitoring gastrointestinal motility in critically ill patients

PURPOSE

The aims of this study were to determine whether bowel sounds auscultation is necessary in critically ill patients and to forecast the prospect of bowel sounds as a monitoring measurement in the intensive care unit (ICU).

BACKGROUND

It has been suggested recently that bowel sounds are not an objective indicator of bowel motility and that auscultation should be abandoned. This has led to confusion as to whether bowel sounds auscultation should be continued in the ICU.

DESCRIPTION OF THE PROJECT

A literature review of articles about bowel sounds and monitoring gastrointestinal motility in critically ill patients was conducted.

OUTCOME

At present, there are no more suitable indicators for bedside monitoring of bowel function and motility than bowel sounds. Although they lack objectivity, bowel sounds give a lot of useful information about gastrointestinal motility. The problems are how to improve practice and assessment standards and enhance the precision of auscultation devices.

CONCLUSION

Bowel sounds auscultation is necessary in the ICU. Effective application in critically ill patients requires reasonable practice and precise instrumentation.

The risk of immune thrombocytopenic purpura after vaccination in children and adolescents

BACKGROUND

The risk of immune thrombocytopenic purpura (ITP) after childhood vaccines other than measles-mumps-rubella vaccine (MMR) is unknown.

METHODS

Using data from 5 managed care organizations for 2000 to 2009, we identified a cohort of 1.8 million children ages 6 weeks to 17 years. Potential ITP cases were identified by using diagnostic codes and platelet counts. All cases were verified by chart review. Incidence rate ratios were calculated comparing the risk of ITP in risk (1 to 42 days after vaccination) and control periods.

RESULTS

There were 197 chart-confirmed ITP cases out of 1.8 million children in the cohort. There was no elevated risk of ITP after any vaccine in early childhood other than MMR in the 12- to 19-month age group. There was a significantly elevated risk of ITP after hepatitis A vaccine at 7 to 17 years of age, and for varicella vaccine and tetanus-diphtheria-acellular pertussis vaccine at 11 to 17 years of age. For hepatitis A, varicella, and tetanus-diphtheria-acellular pertussis vaccines, elevated risks were based on one to two vaccine-exposed cases. Most cases were acute and mild with no long-term sequelae.

CONCLUSIONS

ITP is unlikely after early childhood vaccines other than MMR. Because of the small number of exposed cases and potential confounding, the possible association of ITP with hepatitis A, varicella, and tetanus-diphtheria-acellular pertussis vaccines in older children requires further investigation.

Extracting Hot spots of Topics from Time Stamped Documents

Identifying time periods with a burst of activities related to a topic has been an important problem in analyzing time-stamped documents. In this paper, we propose an approach to extract a hot spot of a given topic in a time-stamped document set. Topics can be basic, containing a simple list of keywords, or complex. Logical relationships such as and, or, and not are used to build complex topics from basic topics. A concept of presence measure of a topic based on fuzzy set theory is introduced to compute the amount of information related to the topic in the document set. Each interval in the time period of the document set is associated with a numeric value which we call the discrepancy score. A high discrepancy score indicates that the documents in the time interval are more focused on the topic than those outside of the time interval. A hot spot of a given topic is defined as a time interval with the highest discrepancy score. We first describe a naive implementation for extracting hot spots. We then construct an algorithm called EHE (Efficient Hot Spot Extraction) using several efficient strategies to improve performance. We also introduce the notion of a topic DAG to facilitate an efficient computation of presence measures of complex topics. The proposed approach is illustrated by several experiments on a subset of the TDT-Pilot Corpus and DBLP conference data set. The experiments show that the proposed EHE algorithm significantly outperforms the naive one, and the extracted hot spots of given topics are meaningful.

Development and evaluation of a data-adaptive alerting algorithm for univariate temporal biosurveillance data

This paper discusses further advances in making robust predictions with the Holt-Winters forecasts for a variety of syndromic time series behaviors and introduces a control-chart detection approach based on these forecasts. Using three collections of time series data, we compare biosurveillance alerting methods with quantified measures of forecast agreement, signal sensitivity, and time-to-detect. The study presents practical rules for initialization and parameterization of biosurveillance time series. Several outbreak scenarios are used for detection comparison. We derive an alerting algorithm from forecasts using Holt-Winters-generalized smoothing for prospective application to daily syndromic time series. The derived algorithm is compared with simple control-chart adaptations and to more computationally intensive regression modeling methods. The comparisons are conducted on background data from both authentic and simulated data streams. Both types of background data include time series that vary widely by both mean value and cyclic or seasonal behavior. Plausible, simulated signals are added to the background data for detection performance testing at signal strengths calculated to be neither too easy nor too hard to separate the compared methods. Results show that both the sensitivity and the timeliness of the Holt-Winters-based algorithm proved to be comparable or superior to that of the more traditional prediction methods used for syndromic surveillance.

A flexibly shaped space-time scan statistic for disease outbreak detection and monitoring

Background

Early detection of disease outbreaks enables public health officials to implement disease control and prevention measures at the earliest possible time. A time periodic geographical disease surveillance system based on a cylindrical space-time scan statistic has been used extensively for disease surveillance along with the SaTScan software. In the purely spatial setting, many different methods have been proposed to detect spatial disease clusters. In particular, some spatial scan statistics are aimed at detecting irregularly shaped clusters which may not be detected by the circular spatial scan statistic.

Results

Based on the flexible purely spatial scan statistic, we propose a flexibly shaped space-time scan statistic for early detection of disease outbreaks. The performance of the proposed space-time scan statistic is compared with that of the cylindrical scan statistic using benchmark data. In order to compare their performances, we have developed a space-time power distribution by extending the purely spatial bivariate power distribution. Daily syndromic surveillance data in Massachusetts, USA, are used to illustrate the proposed test statistic.

Conclusion

The flexible space-time scan statistic is well suited for detecting and monitoring disease outbreaks in irregularly shaped areas.

Modeling and detection of respiratory-related outbreak signatures

Background

Time series methods are commonly used to detect disease outbreak signatures (e.g., signals due to influenza outbreaks and anthrax attacks) from varying respiratory-related diagnostic or syndromic data sources. Typically this involves two components: (i) Using time series methods to model the baseline background distribution (the time series process that is assumed to contain no outbreak signatures), (ii) Detecting outbreak signatures using filter-based time series methods.

Methods

We consider time series models for chest radiograph data obtained from Midwest children's emergency departments. These models incorporate available covariate information such as patient visit counts and smoothed ambient temperature series, as well as time series dependencies on daily and weekly seasonal scales. Respiratory-related outbreak signature detection is based on filtering the one-step-ahead prediction errors obtained from the time series models for the respiratory-complaint background.

Results

Using simulation experiments based on a stochastic model for an anthrax attack, we illustrate the effect of the choice of filter and the statistical models upon radiograph-attributed outbreak signature detection.

Conclusion

We demonstrate the importance of using seasonal autoregressive integrated average time series models (SARIMA) with covariates in the modeling of respiratory-related time series data. We find some homogeneity in the time series models for the respiratory-complaint backgrounds across the Midwest emergency departments studied. Our simulations show that the balance between specificity, sensitivity, and timeliness to detect an outbreak signature differs by the emergency department and the choice of filter. The linear and exponential filters provide a good balance.