Gumbel based p-value approximations for spatial scan statistics

Alcohol consumption: context and association with mortality in Switzerland

PURPOSE

Non-communicable diseases generate the largest number of avoidable deaths often caused by risk factors such as alcohol, smoking, and unhealthy diets. Our study investigates the association between amount and context of alcohol consumption and mortality from major non-communicable diseases in Switzerland.

METHODS

Generalized linear regression models were fitted on data of the cross-sectional population-based National Nutrition Survey menuCH (2014-2015, n = 2057). Mortality rates based on the Swiss mortality data (2015-2018) were modeled by the alcohol consumption group considering the amount and context (i.e., during or outside mealtime) of alcohol consumption and potential confounders. The models were checked for spatial autocorrelation using Moran's I statistic. Integrated nested Laplace approximation (INLA) models were fitted when evidence for missing spatial information was found.

RESULTS

Higher mortality rates were detected among drinkers compared to non-drinkers for all-cancer (rate ratio (RR) ranging from 1.01 to 1.07) and upper aero-digestive tract cancer (RR ranging from 1.15 to 1.20) mortality. Global Moran's I statistic revealed spatial autocorrelation at the Swiss district level for all-cancer mortality. An INLA model led to the identification of three districts with a significant decrease and four districts with a significant increase in all-cancer mortality.

CONCLUSION

Significant associations of alcohol consumption with all-cancer and upper aero-digestive tract cancer mortality were detected. Our study results indicate the need for further studies to improve the next alcohol-prevention scheme and to lower the number of avoidable deaths in Switzerland.

Is Following a Cancer-Protective Lifestyle Linked to Reduced Cancer Mortality Risk?

Objectives: This study investigates the association between a cancer protective lifestyle (defined based on the revised World Cancer Research Fund (WCRF) and the American Institute for Cancer Research (AICR) cancer prevention recommendations) and mortality in Switzerland. Methods: Based on the cross-sectional, population-based National Nutrition Survey, menuCH (n = 2057), adherence to the WCRF/AICR recommendations was assessed via a score. Quasipoisson regression models were fitted to examine the association of adherence to the WCRF/AICR recommendations with mortality at the Swiss district-level. Spatial autocorrelation was tested with global Moran's I. Integrated nested Laplace approximation models were fitted when significant spatial autocorrelation was detected. Results: Participants with higher cancer prevention scores had a significant decrease in all-cause (relative risk 0.95; 95% confidence interval 0.92, 0.99), all-cancer (0.93; 0.89, 0.97), upper aero-digestive tract cancer (0.87; 0.78, 0.97), and prostate cancer (0.81; 0.68, 0.94) mortality, compared to those with lower scores. Conclusion: The inverse association between adherence to the WCRF/AICR recommendations and mortality points out the potential of the lifestyle recommendations to decrease mortality and especially the burden of cancer in Switzerland.

Spatio-temporal epidemiology of the tuberculosis incidence rate in Iran 2008 to 2018

BACKGROUND

Effective reduction of tuberculosis (TB) requires information on the distribution of TB incidence rate across time and location. This study aims to identify the spatio-temporal pattern of TB incidence rate in Iran between 2008 and 2018.

METHODS

This cross-sectional study was conducted on aggregated TB data (50,500 patients) at the provincial level provided by the Ministry of Health in Iran between 2008 and 2018. The Anselin Local Moran's I and Getis-Ord Gi* were performed to identify the spatial variations of the disease. Furthermore, spatial scan statistic was employed for purely temporal and spatio-temporal analyses. In all instances, the null hypothesis of no clusters was rejected at p ≤ 0.05.

RESULTS

The overall incidence rate of TB decreased from 13.46 per 100,000 (95% CI: 13.19-13.73) in 2008 to 10.88 per 100,000 (95% CI: 10.65-11.11) in 2018. The highest incidence rate of TB was observed in southeast and northeast of Iran for the whole study period. Additionally, spatial cluster analysis discovered Khuzestan Province, in the West of the country, having significantly higher rates than neighbouring provinces in terms of both total TB and smear-positive pulmonary TB (SPPTB). Purely temporal analysis showed that high-rate and low-rate clusters were predominantly distributed in the time periods 2010-2014 and 2017-2018. Spatio-temporal results showed that the statistically significant clusters were mainly distributed from centre to the east during the study period. Some high-trend TB and SPPTB statistically significant clusters were found.

CONCLUSION

The results provided an overview of the latest TB spatio-temporal status In Iran and identified decreasing trends of TB in the 2008-2018 period. Despite the decreasing incidence rate, there is still need for screening, and targeting of preventive interventions, especially in high-risk areas. Knowledge of the spatio-temporal pattern of TB can be useful for policy development as the information regarding the high-risk areas would contribute to the selection of areas needed to be targeted for the expansion of health facilities.

A systematic review of aberration detection algorithms used in public health surveillance

The algorithms used for detecting anomalies have evolved substantially over the last decade to take advantage of advances in informatics and to accommodate changes in surveillance data. We identified 145 studies since 2007 that evaluated statistical methods used to detect aberrations in public health surveillance data. For each study, we classified the analytic methods and reviewed the evaluation metrics. We also summarized the practical usage of the detection algorithms in public health surveillance systems worldwide. Traditional methods (e.g., control charts, linear regressions) were the focus of most evaluation studies and continue to be used commonly in practice. There was, however, an increase in the number of studies using forecasting methods and studies applying machine learning methods, hidden Markov models, and Bayesian framework to multivariate datasets. Evaluation studies demonstrated improved accuracy with more sophisticated methods, but these methods do not appear to be used widely in public health practice.

Detecting Drop-offs in Electronic Laboratory Reporting for Communicable Diseases in New York City

CONTEXT

The Bureau of Communicable Disease at the New York City Department of Health and Mental Hygiene receives an average of more than 1000 reports daily via electronic laboratory reporting. Rapid recognition of any laboratory reporting drop-off of test results for 1 or more diseases is necessary to avoid delays in case investigation and outbreak detection.

PROGRAM

We modified our outbreak detection approach using the prospective space-time permutation scan statistic in SaTScan. Instead of searching for spatiotemporal clusters of high case counts, we reconceptualized "space" as "laboratory" and instead searched for clusters of recent low reporting, overall and for each of 52 diseases and 10 hepatitis test types, within individual laboratories. Each analysis controlled for purely temporal trends affecting all laboratories and accounted for multiple testing.

IMPLEMENTATION

A SAS program automatically created input files, invoked SaTScan, and further processed SaTScan analysis results and output summaries to a secure folder. Analysts reviewed output weekly and reported concerning drop-offs to coordinators, who liaised with reporting laboratory staff to investigate and resolve issues.

EVALUATION

During a 42-week evaluation period, October 2017 to July 2018, we detected 62 unique signals of reporting drop-offs. Of these, 39 (63%) were verified as true drop-offs, including failures to generate or transmit files and programming errors. For example, a hospital laboratory stopped reporting influenza after changing a multiplex panel result from "positive" to "detected." Six drop-offs were detected despite low numbers of expected reports missing (<10 per drop-off).

DISCUSSION

Our novel application of SaTScan identified a manageable number of possible electronic laboratory reporting drop-offs for investigation. Ongoing maintenance requirements are minimal but include accounting for laboratory mergers and referrals. Automated analyses facilitated rapid identification and correction of electronic laboratory reporting errors, even with small numbers of expected reports missing, suggesting that our approach might be generalizable to smaller jurisdictions.

Global Research on Syndromic Surveillance from 1993 to 2017: Bibliometric Analysis and Visualization

Syndromic Surveillance aims at analyzing medical data to detect clusters of illness or forecast disease outbreaks. Although the research in this field is flourishing in terms of publications, an insight of the global research output has been overlooked. This paper aims at analyzing the global scientific output of the research from 1993 to 2017. To this end, the paper uses bibliometric analysis and visualization to achieve its goal. Particularly, a data processing framework was proposed based on citation datasets collected from Scopus and Clarivate Analytics’ Web of Science Core Collection (WoSCC). The bibliometric method and Citespace were used to analyze the institutions, countries, and research areas as well as the current hotspots and trends. The preprocessed dataset includes 14,680 citation records. The analysis uncovered USA, England, Canada, France and Australia as the top five most productive countries publishing about Syndromic Surveillance. On the other hand, at the Pinnacle of academic institutions are the US Centers for Disease Control and Prevention (CDC). The reference co-citation analysis uncovered the common research venues and further analysis of the keyword cooccurrence revealed the most trending topics. The findings of this research will help in enriching the field with a comprehensive view of the status and future trends of the research on Syndromic Surveillance.

Evaluation of the Gini Coefficient in Spatial Scan Statistics for Detecting Irregularly Shaped Clusters

Spatial scan statistics with circular or elliptic scanning windows are commonly used for cluster detection in various applications, such as the identification of geographical disease clusters from epidemiological data. It has been pointed out that the method may have difficulty in correctly identifying non-compact, arbitrarily shaped clusters. In this paper, we evaluated the Gini coefficient for detecting irregularly shaped clusters through a simulation study. The Gini coefficient, the use of which in spatial scan statistics was recently proposed, is a criterion measure for optimizing the maximum reported cluster size. Our simulation study results showed that using the Gini coefficient works better than the original spatial scan statistic for identifying irregularly shaped clusters, by reporting an optimized and refined collection of clusters rather than a single larger cluster. We have provided a real data example that seems to support the simulation results. We think that using the Gini coefficient in spatial scan statistics can be helpful for the detection of irregularly shaped clusters.

Using Gini coefficient to determining optimal cluster reporting sizes for spatial scan statistics

BACKGROUND

Spatial and space-time scan statistics are widely used in disease surveillance to identify geographical areas of elevated disease risk and for the early detection of disease outbreaks. With a scan statistic, a scanning window of variable location and size moves across the map to evaluate thousands of overlapping windows as potential clusters, adjusting for the multiple testing. Almost always, the method will find many very similar overlapping clusters, and it is not useful to report all of them. This paper proposes to use the Gini coefficient to help select which of the many overlapping clusters to report.

METHODS

The Gini coefficient provides a quick and intuitive way to evaluate the degree of the heterogeneity of the collection of clusters, which is useful to explain how well the cluster collection reveal the underlying true cluster patterns. Using simulation studies and real cancer mortality data, it is compared with the traditional approach for reporting non-overlapping clusters.

RESULTS

The Gini coefficient can identify a more refined collection of non-overlapping clusters to report. For example, it is able to determine when it makes more sense to report a collection of smaller non-overlapping clusters versus a single large cluster containing all of them. It also fulfils a set of desirable theoretical properties, such as being invariant under a uniform multiplication of the population numbers by the same constant.

CONCLUSIONS

The Gini coefficient can be used to determine which set of non-overlapping clusters to report. It has been implemented in the free SaTScan™ software version 9.3 ( www.satscan.org ).

Simulation-Based Evaluation of the Performances of an Algorithm for Detecting Abnormal Disease-Related Features in Cattle Mortality Records

We performed a simulation study to evaluate the performances of an anomaly detection algorithm considered in the frame of an automated surveillance system of cattle mortality. The method consisted in a combination of temporal regression and spatial cluster detection which allows identifying, for a given week, clusters of spatial units showing an excess of deaths in comparison with their own historical fluctuations. First, we simulated 1,000 outbreaks of a disease causing extra deaths in the French cattle population (about 200,000 herds and 20 million cattle) according to a model mimicking the spreading patterns of an infectious disease and injected these disease-related extra deaths in an authentic mortality dataset, spanning from January 2005 to January 2010. Second, we applied our algorithm on each of the 1,000 semi-synthetic datasets to identify clusters of spatial units showing an excess of deaths considering their own historical fluctuations. Third, we verified if the clusters identified by the algorithm did contain simulated extra deaths in order to evaluate the ability of the algorithm to identify unusual mortality clusters caused by an outbreak. Among the 1,000 simulations, the median duration of simulated outbreaks was 8 weeks, with a median number of 5,627 simulated deaths and 441 infected herds. Within the 12-week trial period, 73% of the simulated outbreaks were detected, with a median timeliness of 1 week, and a mean of 1.4 weeks. The proportion of outbreak weeks flagged by an alarm was 61% (i.e. sensitivity) whereas one in three alarms was a true alarm (i.e. positive predictive value). The performances of the detection algorithm were evaluated for alternative combination of epidemiologic parameters. The results of our study confirmed that in certain conditions automated algorithms could help identifying abnormal cattle mortality increases possibly related to unidentified health events.

p-value approximations for spatial scan statistics using extreme value distributions

Spatial scan statistics are widely applied to identify spatial clusters in geographic disease surveillance. To evaluate the statistical significance of detected clusters, Monte Carlo hypothesis testing is often used because the null distribution of spatial scan statistics is not known. A drawback of the method is that we have to increase the number of replications to obtain accurate p-values. Gumbel-based p-value approximations for spatial scan statistics have recently been proposed and evaluated for Poisson and Bernoulli models. In this study, we examine the use of a generalized extreme value distribution to approximate the null distribution of spatial scan statistics as well as the Gumbel distribution. Through simulation, p-value approximations using extreme value distributions for spatial scan statistics are assessed for multinomial and ordinal models in addition to Poisson and Bernoulli models.

Spatial epidemiology of Escherichia coli O157:H7 in dairy cattle in relation to night roosts Of Sturnus vulgaris (European Starling) in Ohio, USA (2007-2009)

The goal of our study was to use spatial scan statics to determine whether the night roosts of European starlings (Sturnus vulgaris) act as point sources for the dissemination of Escherichia coli O157:H7 among dairy farms. From 2007 to 2009, we collected bovine faecal samples (n = 9000) and starling gastrointestinal contents (n = 430) from 150 dairy farms in northeastern Ohio, USA. Isolates of E. coli O157:H7 recovered from these samples were subtyped using multilocus variable-number tandem repeat analysis (MLVA). Generated MLVA types were used to construct a dendrogram based on a categorical multistate coefficient and unweighted pair-group method with arithmetic mean (UPGMA). Using a focused spatial scan statistic, we identified statistically significant spatial clusters among dairy farms surrounding starling night roosts, with an increased prevalence of E. coli O157:H7-positive bovine faecal pats, increased diversity of distinguishable MLVA types and a greater number of isolates with MLVA types from bovine-starling clades versus bovine-only clades. Thus, our findings are compatible with the hypothesis that starlings have a role in the dissemination of E. coli O157:H7 among dairy farms, and further research into starling management is warranted.

Maximum linkage space-time permutation scan statistics for disease outbreak detection

BACKGROUND

In disease surveillance, the prospective space-time permutation scan statistic is commonly used for the early detection of disease outbreaks. The scanning window that defines potential clusters of diseases is cylindrical in shape, which does not allow incorporating into the cluster shape potential factors that can contribute to the spread of the disease, such as information about roads, landscape, among others. Furthermore, the cylinder scanning window assumes that the spatial extent of the cluster does not change in time. Alternatively, a dynamic space-time cluster may indicate the potential spread of the disease through time. For instance, the cluster may decrease over time indicating that the spread of the disease is vanishing.

METHODS

This paper proposes two irregularly shaped space-time permutation scan statistics. The cluster geometry is dynamically created using a graph structure. The graph can be created to include nearest-neighbor structures, geographical adjacency information or any relevant prior information regarding the contagious behavior of the event under surveillance.

RESULTS

The new methods are illustrated using influenza cases in three New England states, and compared with the cylindrical version. A simulation study is provided to investigate some properties of the proposed arbitrary cluster detection techniques.

CONCLUSION

We have successfully developed two new space-time permutation scan statistics methods with irregular shapes and improved computational performance. The results demonstrate the potential of these methods to quickly detect disease outbreaks with irregular geometries. Future work aims at performing intensive simulation studies to evaluate the proposed methods using different scenarios, number of cases, and graph structures.

Event detection using Twitter: a spatio-temporal approach

BACKGROUND

Every day, around 400 million tweets are sent worldwide, which has become a rich source for detecting, monitoring and analysing news stories and special (disaster) events. Existing research within this field follows key words attributed to an event, monitoring temporal changes in word usage. However, this method requires prior knowledge of the event in order to know which words to follow, and does not guarantee that the words chosen will be the most appropriate to monitor.

METHODS

This paper suggests an alternative methodology for event detection using space-time scan statistics (STSS). This technique looks for clusters within the dataset across both space and time, regardless of tweet content. It is expected that clusters of tweets will emerge during spatio-temporally relevant events, as people will tweet more than expected in order to describe the event and spread information. The special event used as a case study is the 2013 London helicopter crash.

RESULTS AND CONCLUSION

A spatio-temporally significant cluster is found relating to the London helicopter crash. Although the cluster only remains significant for a relatively short time, it is rich in information, such as important key words and photographs. The method also detects other special events such as football matches, as well as train and flight delays from Twitter data. These findings demonstrate that STSS is an effective approach to analysing Twitter data for event detection.

Applying spatial analysis tools in public health: an example using SaTScan to detect geographic targets for colorectal cancer screening interventions

Epidemiologists are gradually incorporating spatial analysis into health-related research as geocoded cases of disease become widely available and health-focused geospatial computer applications are developed. One health-focused application of spatial analysis is cluster detection. Using cluster detection to identify geographic areas with high-risk populations and then screening those populations for disease can improve cancer control. SaTScan is a free cluster-detection software application used by epidemiologists around the world to describe spatial clusters of infectious and chronic disease, as well as disease vectors and risk factors. The objectives of this article are to describe how spatial analysis can be used in cancer control to detect geographic areas in need of colorectal cancer screening intervention, identify issues commonly encountered by SaTScan users, detail how to select the appropriate methods for using SaTScan, and explain how method selection can affect results. As an example, we used various methods to detect areas in Florida where the population is at high risk for late-stage diagnosis of colorectal cancer. We found that much of our analysis was underpowered and that no single method detected all clusters of statistical or public health significance. However, all methods detected 1 area as high risk; this area is potentially a priority area for a screening intervention. Cluster detection can be incorporated into routine public health operations, but the challenge is to identify areas in which the burden of disease can be alleviated through public health intervention. Reliance on SaTScan’s default settings does not always produce pertinent results.

New developments in the spatial scan statistic

The quantity and variety of spatial data have increased over recent years, and the variety and sophistication of tools for analysing this type of data have also increased. One such tool is the spatial scan statistic, which is freely available ( www.satscan.org ) and has been the subject of much scholarly research since its introduction in 1995 owing to its numerous applications in epidemiology, criminology and other fields. This paper provides readers with a non-technical introduction to the spatial scan statistic, together with an overview of associated research, which focuses particularly on work conducted at the University of Sheffield’s Information School, in collaboration with the School of Health and Related Research. This work falls into three main areas. First, we provide an examination of the probability of obtaining false alerts when using the statistic, and ways in which this can be managed. Second, we describe the development of a definitive way of measuring the spatial accuracy of the statistic. Third, and potentially the most important in terms of impact, we discuss a means of substantially increasing the detection capability of the statistic by placing a realistic constraint on the strength of any cluster that is likely to be present in the data. The paper also provides a discussion of potential future research directions.

Data-driven inference for the spatial scan statistic

Background

Kulldorff's spatial scan statistic for aggregated area maps searches for clusters of cases without specifying their size (number of areas) or geographic location in advance. Their statistical significance is tested while adjusting for the multiple testing inherent in such a procedure. However, as is shown in this work, this adjustment is not done in an even manner for all possible cluster sizes.

Results

A modification is proposed to the usual inference test of the spatial scan statistic, incorporating additional information about the size of the most likely cluster found. A new interpretation of the results of the spatial scan statistic is done, posing a modified inference question: what is the probability that the null hypothesis is rejected for the original observed cases map with a most likely cluster of size k, taking into account only those most likely clusters of size k found under null hypothesis for comparison? This question is especially important when the p-value computed by the usual inference process is near the alpha significance level, regarding the correctness of the decision based in this inference.

Conclusions

A practical procedure is provided to make more accurate inferences about the most likely cluster found by the spatial scan statistic.

Local multiplicity adjustment for the spatial scan statistic using the Gumbel distribution

The spatial scan statistic is an important and widely used tool for cluster detection. It is based on the simultaneous evaluation of the statistical significance of the maximum likelihood ratio test statistic over a large collection of potential clusters. In most cluster detection problems, there is variation in the extent of local multiplicity across the study region. For example, using a fixed maximum geographic radius for clusters, urban areas typically have many overlapping potential clusters, whereas rural areas have relatively few. The spatial scan statistic does not account for local multiplicity variation. We describe a previously proposed local multiplicity adjustment based on a nested Bonferroni correction and propose a novel adjustment based on a Gumbel distribution approximation to the distribution of a local scan statistic. We compare the performance of all three statistics in terms of power and a novel unbiased cluster detection criterion. These methods are then applied to the well-known New York leukemia dataset and a Wisconsin breast cancer incidence dataset.