Summer school: Bayesian methods in health economics

Bayesian methods in health economics

20-24 July 2020, Centro Studi CISL, Florence (Italy)

Instructors: Gianluca Baio (UCL), Anna Heath (SickKids hospital, Toronto), Chris Jackson (MRC Biostatistics Unit), Mark Strong (Sheffield), Nicky Welton (Bristol) & Howard Thom (Bristol)

This residential summer school aims at providing an introduction to Bayesian analysis and Markov Chain Monte Carlo (MCMC) methods using R and MCMC sampling software (such as OpenBUGS and JAGS), as applied to cost-effectiveness analysis and typical models used in health economic evaluations. We will also focus on more recent methods for Probabilistic Sensitivity Analysis including Value of Information calculations. As such, it is intended for health economists, statisticians, and decision modellers interested in the practice of Bayesian modelling and will be based on a mixture of lectures and computer practicals. The emphasis will be on examples of applied analysis: software and code to carry out the analyses will be provided.

Participants are encouraged to bring their own laptops for the practicals. We shall assume a basic knowledge of standard methods in health economics and some familiarity with a range of probability distributions, regression analysis, Markov models and random-effects meta-analysis. However, statistical concepts are reviewed in the context of applied health economic evaluations in the lectures. Lectures and practicals are based on Bayesian Methods in Health Economics (BMHE), The BUGS Book (BB) and Evidence Synthesis for Decision Making in Healthcare (ESDM), which will all be provided as part of the course material (ie included in the registration fee).

The summer school is hosted in the beautiful complex of the Centro Studi CISL, overlooking and a short distance from Florence. The registration fees include full board accommodation in the Centro Studi.

Lecture topics

  1. Introduction to health economic evaluations
  2. Introduction to Bayesian inference
  3. Introduction to Markov Chain Monte Carlo in BUGS
  4. Cost-effectiveness analysis with individual-level data
  5. Aggregated-level data and hierarchical models
  6. Evidence synthesis and network meta-analysis
  7. Model error and structural uncertainty
  8. Markov models
  9. Survival analysis
  10. Missing data in cost-effectiveness modelling
  11. Introduction to the value of information
  12. Expected value of partial information (1) - algebraic tricks
  13. Expected value of partial information (2) - generalised additive models & GP regression
  14. Expected value of partial information (3) - GP regression in INLA/SPDE
  15. Expected value of sample information (1) - conjugated analysis
  16. Expected value of sample information (2) - efficient nested simulation and moment matching
  17. Expected value of sample information (3) - regression- and sufficient statistics-based methods

Software & useful information

  • OpenBUGS (free software for Bayesian analysis)
  • R (free general statistical software)
  • JAGS (alternative software for Bayesian analysis) - probably the easiest option for Linux or Mac users
  • Stan (yet another software for Bayesian analysis) - this is based on a different method for MCMC (called Hamiltonian Monte Carlo)
  • R2OpenBUGS (R library to interface R and OpenBUGS) or R2jags (does the same for R and JAGS) or rstan (does the same for R and Stan)
  • BCEA (R library to perform Bayesian Cost Effectiveness Analysis). The stable (CRAN) current version is 2.5. The GitHub version is 2.6.
  • Wine (a “compatibility layer” that allows to run Windows applications from Linux or Mac)
  • Instructions to install OpenBUGS using Wine (for Mac users)

How to get there

The course will be held in the large seminar room at the Centro Studi CISL, Via della Piazzuola, 71 — 50133 Florence. The registration fee includes accommodation for the duration of the course (from Sunday night to Thursday night included), but specific arrangements for extra days could be made directly with the Centro Studi, pending room availabilities.

Florence is served by the Amerigo Vespucci international airport, but the most convenient way to fly to Tuscany is probably through Pisa’s Galileo Galilei international airport — there are some low cost airlines flying there and the airport is a bit bigger. There is a train connection from Pisa Airport to Pisa Centrale and then to Firenze Santa Maria Novella. There are also coaches going directly to Florence from Pisa airport – some options here. Train tickets can be booked and bought at the Trenitalia website. If you arrive to Florence Airport, the taxi ride to the Centro Studi is about 20 minutes and cost about €20. There is also a recently opened tram connection from Florence airport to Santa Maria Novella train station.

Another relatively convenient airport is Bologna (which is just half an hour away on the fast, but more expensive train, called “Frecciarossa”, still available to book from Trenitalia. There’s a bus connection from Bologna airport to the main train station and then lots of connections from Bologna Centrale to Firenze Santa Maria Novella). Rome is also not too far away. There are two main train companies in Italy: Trenitalia is state-run while Italo is a private company. Only Trenitalia serves Florence from Pisa, while it is possible to travel to Florence from other cities with both.

Once you arrive to Florence main station, you could get to Centro Studi either by taxi (there is a taxi queue in front of the station), or by bus. You will have to buy your bus ticket before you board the bus — you can buy a 90-minute ticket for €1.50, but other options are available). From Santa Maria Novella, you can reach Centro Studi with the following:

  • Bus no 7 to Fiesole (approximately one every 20 minutes) and get off at Ospedale di Camerata. Via Della Piazzola (where Centro Studi is) runs just parallel to Ospedale di Camerata about 300 meters away.

This is the official route suggested on Centro Studi’s website, but for example Google Maps also suggest other options. If you are driving, there is parking space in the Centro Studi, but it is probably best to arrange directly with them.

Places to go in Florence (et al)

Not that you will want to ever leave the beatiful Centro Studi and in fact usually we need to find quiet places to escape the participants who want to keep talking shop even when we’re dying to enjoy a drink in the terrace… But if you fancy going out for dinner, here are some options (as somebody from Florence, but who hasn’t lived there for way longer than he cares to remember…)

  • Mercato Centrale. That’s the old market — you can find stall selling all sorts of things (mainly leather products). You can also have dinner in the main food hall, where you can find several stores serving pasta, pizza and good food using local products.
  • Pescepane. A fish restaurant in the fancy area of Sant’Ambrogio.
  • Il Rifrullo. A nice cocktail bar where you can also have “aperitivo” (that’s the Italian version of sitting in the terrace, drinking and getting buffet food) in the super cool area of San Niccolò, just below Piazzale Michelangelo.
  • Generally, the area of Santo Spirito. There are plenty of restaurants and bars and locals will flock to get some respite from the hot weather.
  • You can also walk to Fiesole (15 to 20 minutes), to visit the beautiful Roman theatre for the Estate Fiesolana.
  • As somebody from Florence, I really shouldn’t say, but Pisa really is beatiful — so if you are flying to/from Pisa, it’s worth spending some time wandering around the old town. Other amazing day-trips (in case you are coming before/staying after the summer school, that is…) include Siena and Lucca. Not to brag, but basically (almost!) anywhere in Tuscany is awesome…

Registration

There are three types of registration:

  1. Students (£800)
  2. Public sector (£1,500)
  3. Private sector (£1,800)

Registration is open now from the UCL Online Store.

All course fees include full board accommodation, copies of BMHE, BB and ESDM plus course material (slides, R/BUGS codes for the practicals, relevant papers).

Avatar
Gianluca Baio
Professor of Statistics and Health Economics

Related