Bayesian Statistical Analysis

 Bayesian heading

Details of how to book these courses will be available in 2017

Course 1

Bayesian Biostatistics: Special Topics

Course suite 2

Bayesian Disease Mapping

The courses are organised by Dr Pamela Warner, CRFR and Centre for Population Health Sciences (UIPHSI), at The University of Edinburgh.

Course presenter


Both courses are given by Professor Andrew B. Lawson of the Medical University of South Carolina. Professor Lawson has wide ranging experience of the application of Bayesian modeling to Biostatistical applications. He is an elected fellow of the American Statistical Association and author of many Biostatistical journal papers and a range of books on Bayesian methods in disease mapping and more general Biostatistical application areas. He has run many courses on related topics over a number of years, across the world.

Course 1

Overview – Bayesian Biostatistics: Special Topics

This two day course is designed for those with some previous experience of Bayesian statistical methods but who want to extend their experience into a variety of specialist areas of Biostatistical applications. The course provides coverage of the use of Bayesian software (WinBUGS/OpenBUGS and INLA) in the analysis of topic areas: linear models and Bayesian GLMMs, longitudinal analysis, Survival analysis, Meta analysis, Bioassay, Measurement Error, Imaging, and Disease mapping.

For this course a copy of the recent text by Lesaffre, E. and Lawson, A. B. (2012) Bayesian Biostatistics, Wiley, New York  will be included in the course fee. Material included in the course is covered in the second half of the text and many examples are provided from those covered in the book.

Who should attend?

The course focusses on special applications found commonly in Biostatistics and is designed for those who already have some experience of applying Bayesian methods but want to extend their experience to a wider array of software (e.g. INLA) and a wider array of topics commonly found in Biostatistical collaborations.

Programme course 1

Day 1

Bayesian methods introduction
  • Intro with 2 group examples and shown results from WB and INLA
WinBUGS/OpenBUGS and INLA11-12.30pm
  • Hands on with simple examples in WB and INLA
1.30 – 3pm
Regression, Linear mixed models (LMMs), GLMMs (1hr)
Variable selection (1/2hr)
  • Regression topics
3.30 – 5pm
Longitudinal analysis incl missingness (1hr)
Survival (1/2hr)
  • Basic longitudinal
    and survival modeling

Day 2

  • Measurement Error (ME), SEMs and instrumental variables
11.00 – 12pm
  • Meta-analysis
1 – 2.30pm
  • Imaging (1hr)
  • Bioassay (1/2hr)
  • Disease Mapping

Course suite 2

Overview – Bayesian Disease Mapping suite of courses

Bayesian Disease Mapping comprises 3 hands-on spatial statistical modelling courses.

Bayesian Disease Mapping is split into three courses: a course each at Introductory and Advanced levels (IBDM, ABDM), plus a new addition in 2014, an Introductory course on INLA (BDMI: Bayesian Disease Mapping with INLA).

These three courses can be booked singly or in combination. Understanding/expertise to the level of IBDM (Introductory course) is required/advised for attendance at either of the other two courses and, in addition, BDMI (INLA software) skill-set is a pre-requisite for attending the ABDM (Advanced) course. Discounts apply for those from Academic/Charity institutions, and when a combination of two or more of the suite courses is booked by one individual.

Participants will gain an in-depth understanding of the basic issues, methods and techniques used in the analysis of spatial health data using a Bayesian approach. They will gain insight into the detailed analysis of practical problems in risk estimation and cluster detection. The courses are presented by a leading researcher in the field of disease mapping and spatial epidemiology.

Who should attend?

The courses are intended for epidemiologists and public health workers who need to analyse geographical disease incidence. In addition, the courses may be of interest to statisticians or geographers and planners who deal with spatial disease data. Some statistical/epidemiological background would be beneficial but is not essential.

Programme for course suite 2

These courses are designed to provide a comprehensive introduction to the area of Bayesian disease mapping in applications to Public Health and Epidemiology.
Introductory Module (IBDM)


IBDM  is designed for those who wish to gain a basic insight into the methods of Bayesian Disease Mapping, with introduction to small area statistical and epidemiological issues and basic spatial modelling using WinBUGS. A review of alternative BDM software is also provided.

Booking this course entitles you to a complimentary copy of Prof Lawson’s latest edition of his textbook: Bayesian Disease Mapping: Hierarchical Modeling in Spatial Epidemiology, 2nd Ed (2013)

Day 1
  • Basic concepts of Bayesian methods and disease mapping
  • Bayesian computation and MCMC
  • Basic R and WinBUGS use
  • Demonstration of risk estimation and cluster detection using WinBUGS
Day 2
  • Hands-on with simple WinBUGS models: Poisson-gamma; convolution models for risk estimation
  • Ecological analysis, cluster models
  • Basic Space-Time models: Ohio respiratory cancer; seasonal effects
  • Space-Time Kalman-filtering
  • Use of R2WinBUGS, BRugs, OpenBUGS  and related software
 INLA Basics (BDMI)
BDMI  is designed for those who  want a basic introduction to the use of INLA for Bayesian Disease Mapping. The course is intended for those have had a basic introduction to BDM (such as on the IBDM course) and will cover.

  • Basic use of R and R graphics
  • fitting basic BDM models using INLA.

Some R experience would be useful but not essential

Advanced Module (ABDM)


ABDM  is designed for those who want to cover advanced BDM methods, and includes advanced use of WinBUGS. The course will include theoretical input, but also practical elements and participants will be involved hands-on in the use of R and WinBUGS in disease mapping applications. Both spatial and spatio-temporal analyses will be considered and space-time modelling with INLA will be covered. Examples will range over childhood asthma data from Georgia, influenza in South Carolina, foot-and-mouth disease in the UK and Ohio respiratory cancer.

Day 1
  • Spatial models and simple variants: convolution, proper CAR, full MVN
  • Special application: Case event modelling
  • Special applications: sparse count data: zip and factorial regression
  • Multiple disease analysis
  • Spatial survival modelling
Day 2
  • Measurement error, SEMS and Joint modelling. CPO and pseudo Bayes factor
  • Infectious disease models and veterinary data
  • Regression and variable selection;
  • INLA space-time modelling

This course requires both Introductory level BDM understanding and experience, and facility with the basics of INLA.


Attendees must bring a laptop with R and WinBUGS 1.4.3 software preloaded. Datasets will be provided. R and WinBUGS software can be downloaded from the following websites: and/or and



bayesian book
Bayesian Disease Mapping: Hierarchical Modeling in Spatial Epidemiology, 2nd Ed

Since the publication of the first edition, many new Bayesian tools and methods have been developed for space-time data analysis, the predictive modeling of health outcomes, and other spatial biostatistical areas – hence this 2nd edition of the book.

This book is included as part of the course support materials only for the Introductory Module (taken a a single course or part of a bundle of suite 2 courses).


Developed by Learning Technology Section, University of Edinburgh  |  Privacy Policy
Banner Images: © NHS Scotland, 2011 & ©