Scan Statistics and SaTScan
In statistics, a Scan statistic or window statistic is a problem relating to the clustering of randomly positioned points.
Scan Statistics and SaTScan
Overview of our Scan Statistics and SaTScan Training Programs
Scan statistics are the most important tool for detecting clusters. A scan
statistic has the objective of detecting and evaluating the statistical significance
of clusters that cannot be explained by the assumption of randomness. This
is done by moving a window over the study region and identifying a region, if
there is one, with a higher concentration of points than should occur by chance.
Detection of such regions is traditionally performed by maximizing a likelihood ratio
as will be shown.
Researchers encounter at least nine types of datasets where they might wish
to identify spatial clusters using scan statistics:
* point data
* case-control data
* aggregated data
* spatial-temporal data
* spatial survival data
* event data
* multinomial data
* ordinal data
* time series data
To analyze Scan Statistics data, SaTScan software is essential. It is
a free software that analyzes spatial, temporal and space-time data using the spatial, temporal,
or space-time scan statistics.
It is designed for any of the following interrelated purposes:
* Perform geographical surveillance of disease, to detect spatial or space-time
disease clusters, and to see if they are statistically significant.
* Test whether a disease is randomly distributed over space, over
time or over space and time.
* Evaluate the statistical significance of disease cluster alarms.
* Perform prospective real-time or time-periodic disease surveillance for
the early detection of disease outbreaks.
The overall course have three parts. The first part of the course introduce
Scan-statistics and summarizes useful formulas. The second part of the course is a
step-by-step tutorial for the SaTScan™ software. The last part of the course concentrates
on applications of the different types of scan statistics using SaTScan™ software.
This program is provided by Statistician, Biostatistician,
Epidemiologists, and Researchers.
Tailor-Made Course
We can prepare SaTScan software trainings as tailor-made course to meet individual or organization-wide needs. A training needs assessment will be done on the training participants to collect data on the existing skills, knowledge gaps, training expectations, and tailor-made needs.
Explore Traning Programs
Introduction to Scan Statistics and Methods
This training course introduce Scan-statistics and summarizes useful formulae for approximate and exact probabilities and moments of scan statistics with retrospective and prospective scanning of events over time, and references available tables and computer programs.
Course learning outcomes
After studying this course, you should be able to:
- Understand the types of Scan Statistics
- Differentiate a Retrospective and Prospective Scanning of Events over Time
- Understand Success Scans in a Sequence of Trials
- Understand Higher-Dimensional Scans
Course content
1. Introduction to Scan Statistics
- Definition of Scan Statistics
- Types of Scan Statistics
- Spatial, Temporal and Space-Time Scan Statistics
- Scan Statistic Models
- Bernoulli versus Ordinal Mode
- Normal versus Exponential Model
- Normal versus Ordinal Model
- Discrete versus Homogeneous Poisson Model
- The Discrete Scan Statistics
- One-Dimensional Case
- Multidimensional Case
- The Continuous Scan Statistics
- One-Dimensional Case
- Multidimensional Case
- Power of the Scan Statistic
- Clusters and Intuition
- Applications of Scan Statistic for Medical and Health Sciences
2. Retrospective Scanning of Events Over Time
- Conditional Case: Uniform Distribution of Events
- The Scan Statistic on the Circle
- The Ratchet Scan Statistic
- Moments of Scan Statistics
- Exact Values for Moments of the Scan Statistic
3. Prospective Scanning of Events over Time
- Poisson Distribution of Events
- Handling Trends or Seasonality in Data
- Moments of Scan Statistics
- The Expected Waiting Time until a Cluster
- The Distribution of the Number of Clusters
- The Scan Statistic on the Circle
4. Success Scans in a Sequence of Trials
- Binomial Distribution of Events: Discrete Time, Unconditional Case
- A Null Model for the Unconditional Case: The Bernoulli Process
- The Charge Problem
- Binomial Distributed Events: Discrete Time, Conditional Case
- Moments of Scan Statistics
5. Higher-Dimensional Scans
- Introduction
- The Conditional Problem
- The Unconditional Problem
- Clustering on the Lattice
6. Scan Statistics in DNA and Protein Sequence Analysis
- Introduction
- Scanning for Clusters of Patterns
- Matching in DNA Sequences
- Matching in Multiple Random Letter Sequences
- Sequencing Fragments to Reconstruct a Genome
- Using Double Scans for More Effective Searches for Homologies
- Correlated Descendant Sequences Scan Statistics
Target Participants
The intended audience for this training activity is Statisticians, Biostatisticians and Researchers including Epidemiologist, Biologists, Clinicians, Laboratory and Health professionals with the following prerequisite.
Prerequisites
No prior working knowledge of SaTScan software is required for this course. Understanding of the basic statistical/ bio-statistical concepts, basic computer operational skills and data intuition are required.
Training Approach
This training on Scan Statistics and Methods is delivered by our seasoned trainers who have vast experience as expert professionals using SaTScan Software. The course is mainly focused on theoretical parts of Scan Statistics and Methods and taught through a mix of practical activities, theory and group works.
Training manuals and additional reference materials are provided to the participants.
Certification
Upon successful completion of this course, participants will be issued with a certificate.
Tailor-Made Course
We can also do this as a tailor-made course to meet organization-wide needs. A training needs assessment will be done on the training participants to collect data on the existing skills, knowledge gaps, training expectations, and tailor-made needs.
The SaTScan Software
SaTScan is a free software that analyzes spatial, temporal and space-time data using the spatial, temporal, or space-time scan statistics.
This training course is a step-by-step training for the SaTScan™ software. This course also demonstrate that analyzes spatial, temporal and space-time data using the spatial, temporal, or space-time scan statistics to detect spatial or space-time disease clusters, and to determine if they are statistically significant. The only prerequisite knowledge is a basic understanding of statistics (Scan-statistic) and epidemiology.
Course learning outcomes
After studying this course, you should be able to:
- Understand how to start SaTScan
- Understand the SaTScan Software application
- Understand the SaTScan Software parameters
- Enter basic data into SaTScan
- Run SaTScan and analyse data
- Interpret SaTScan outputs
Course content
1. Introduction to SaTScan
- The Purpose of SaTScan
- SaTScan with Google Earth
- Data Types and Methods
- SaTScan Software Download and Installation
2. Launching the SaTScan Software
- Input Data
- SaTScan Import Wizard
- Basic SaTScan Features
- Advanced SaTScan Features
- Running SaTScan
- Results of Analysis and Outputs
- KML Geographical Output File
- Shapefile Geographical Output
- Temporal Graphs HTML File
- Cluster Information File
- Stratified Cluster Information File
- Location Information File
- Risk Estimates for Each Location File
- Simulated Log Likelihood Ratios File
3. Covariate Adjustments
- Covariate Adjustment Using the Input Files
- Covariate Adjustment Using Statistical Regression Software
- Covariate Adjustment Using Multiple Data Sets
4. Spatial and Temporal Adjustments
- Adjusting for Temporal Trends
- Adjusting for Day-of-Week Effects
- Adjusting for Purely Spatial Clusters
- Adjusting for Known Relative Risks
5. Handling Missing Data
- Missing Data for some locations and times
Target Participants
The intended audience for this training activity is Statisticians, Biostatisticians and Researchers including Epidemiologist, Biologists, Clinicians, Laboratory and Health professionals with the following prerequisite.
Prerequisites
No prior working knowledge of SaTScan software is required for this course. Understanding of the basic statistical/ bio-statistical concepts, basic computer operational skills and data intuition are required.
Training Approach
This training on SaTScan Software is delivered by our seasoned trainers who have vast experience as expert professionals using SaTScan Software. The course is mainly focused on step-by-step practice for SaTScan Software and taught through a mix of practical activities, theory and group works.
Training manuals and additional reference materials are provided to the participants.
Certification
Upon successful completion of this course, participants will be issued with a certificate.
Tailor-Made Course
We can also do this as a tailor-made course to meet organization-wide needs. A training needs assessment will be done on the training participants to collect data on the existing skills, knowledge gaps, training expectations, and tailor-made needs.
Scan Distribution Theory and its Applications using SaTScan
SaTScan is a free software that analyzes spatial, temporal and space-time data using the spatial, temporal, or space-time scan statistics.
This training course concentrates on applications of the different types of scan statistics using SaTScan™ software. Mainly this part of the course is designed for researchers in many medical and health science, and applied fields, and is written from the point of view of methodology and applications of the scan statistics. Applications of a wide variety of scan statistics to fields ranging from public health, epidemiology, biology, molecular biology, quality control, reliability, and other fields are illustrated. This part of the course also brings together and reviews the statistical and mathematical approaches used to develop exact and approximate results for the scan statistics.
Course learning outcomes
After studying this course, you should be able to:
- Understand the SaTScan Software application
- Run SaTScan and analyse data
- Enter basic data into SaTScan
- Understand and analyse Purely Spatial Scan Statistic
- Understand and analyse Space-Time Scan Statistic
- Understand and analyse Space–Time Permutation Scan Statistic
- Interpret SaTScan outputs
Course content
1. Purely Spatial Scan Statistic
- Introduction
- Poisson and Bernoulli Models
- Likelihood ratio test
- Properties of the Test Statistic
- Detection versus inference
- Power
- Computations and Monte Carlo Sampling
- Limitations of Repeated Purely Spatial Analyses
- Applications
- Illustrate the Purely Spatial Poisson Scan Statistic models using New York State Cancer Incidence Data
- Illustrate the Bernoulli Spatial Scan Statistic models using the New York State Birth Defect Data
2. Space-Time Scan Statistic
- Introduction
- Poisson Model versus Negative Binomial Distribution
- Prospective Poisson Space-time Scan Statistic
- Testing for Clustering: Detecting emerging clusters
- Time periodic surveillance
- Applications
- Illustrate the model and Evaluating Cluster Alarm using data on Brain Cancer Incidence in New Mexico
- Illustrate rapid surveillance and detecting and evaluating emerging clusters using data on COVID-19 patients in Ethiopia
3. Space–Time Permutation Scan Statistic
- Introduction
- Hyper-geometric distribution
- Poisson generalized likelihood ratio
- Monte Carlo hypothesis testing
- Missing Data
- Application
- Prospective Time Periodic Geographical COVID-19 Surveillance in Ethiopia using a Space-Time Scan Statistics: Detecting and Evaluating Emerging Clusters
4. Miscellaneous Models in Scan Statistic
- Ordinal Scan Statistic for Identifying Unusual Cancer Stage Patterns
- Multinomial Scan Statistic for Identifying Unusual Population Age Structures
- The Knox method and other tests for space-time interaction
- Power comparisons for disease clustering tests: Computational Statistics and Data Analysis
- Evaluating disease outbreak detection methods: Benchmark data and power calculations
- Bayesian Network Scan Statistics for Multivariate Pattern Detection
Target Participants
The intended audience for this training activity is Statisticians, Biostatisticians and Researchers including Epidemiologist, Biologists, Clinicians, Laboratory and Health professionals with the following prerequisite.
Prerequisites
No prior working knowledge of SaTScan software is required for this course. Understanding of the basic statistical/ bio-statistical concepts, basic computer operational skills and data intuition are required.
Training Approach
This training on Scan Distribution Theory and its Applications is delivered by our seasoned trainers who have vast experience as expert professionals using SaTScan Software. The course is mainly focused on step-by-step practice for SaTScan Software and taught through a mix of practical activities, theory and group works.
Training manuals and additional reference materials are provided to the participants.
Certification
Upon successful completion of this course, participants will be issued with a certificate.
Tailor-Made Course
We can also do this as a tailor-made course to meet organization-wide needs. A training needs assessment will be done on the training participants to collect data on the existing skills, knowledge gaps, training expectations, and tailor-made needs.
Frequently Asked Questions
When can I start this course?
- This course is open enrollment, so you can register and start the course whenever you are ready.
What happens when I complete the course?
- You will automatically get a certificate of completion as soon as you complete the course and pass the graded quizzes and project.