Statistics for health registers and linked databases

20th to 21th May 2009, The Open University, Milton Keynes

The Open University Statistics Group presented a two-day conference on Statistics for health registers and linked databases, held on Wednesday May 20th – Thursday May 21st, 2009, at the Open University in Milton Keynes. 

In medicine, epidemiology and the social sciences, increasing use is now made of electronic databases of individual records, sometimes augmented by linkage, and often set up for purposes other than health research. The analysis of such data offers huge potential, but also presents great challenges. Statistics for health registers and linked databases, the 20th nearly-annual Open University statistics conference, will focus on the statistical issues and methods relevant to the analysis of such data, including: probabilistic linkage, methods of confounder control, handling missing data, multiplicities and multiple data sources, survival analysis methods, confidentiality. 

We gratefully acknowledge financial support from the Wellcome Trust and The Open University. 

The conference included 17 invited talks, speakers are listed below, abstracts and links to presentations are given on the programme page. We also invited poster presentations on relevant topics. We would like to thank all presenters for contributing to the success of this conference. 

Organising Committee

Paddy Farrington, Sarah Frain, Heather Whitaker, Mario Cortina, Liam Smeeth


General talks
Prof Sheila Bird (MRC Biostatistics Unit, Cambridge)
Prof Paul Boyle (St Andrews University) 

Record linkage and confidentiality
Professor Harvey Goldstein (Bristol University)
Prof Catherine Quantin (Dijon University)
Dr Mario Cortina (Institute of Child Health, London) 

Causal inference
Dr Stijn Vansteelandt (Ghent University) 

Missing data
Dr Bianca De Stavola (London School of Hygiene and Tropical Medicine)
Dr Chris Jackson (MRC Biostatistics Unit, Cambridge)
Prof Marie Reilly (Karolinska Institute, Stockholm) 

Event history analysis
Prof Paddy Farrington and Dr Heather Whitaker (Open University, Milton Keynes)
Prof Hans van Houwelingen (Leiden University)
Dr Bendix Carstensen (Steno Diabetes Centre, Denmark)

Dr Daniela De Angelis (MRC Biostatistics Unit, Cambridge and Health Protection Agency, London)
Mr Ismail Ahmed and Dr Pascale Tubert-Bitter (INSERM, Paris) 

Round-table discussion
Prof Liam Smeeth (London School of Hygiene and Tropical Medicine)
Prof Carol Dezateux (Institute of Child Health, London) 


Wednesday 20th May 
10.15 Registration opens
10.30 Coffee and pastries 

10.55 Conference opens
11.00 Sheila Bird, Best laid plans: record linkage illuminates public health, and other jurisdictions
11.45 Harvey Goldstein, Efficient use of prior information about record mismatches in analysing linked data 

12.15 Lunch + posters 

13.45 Mario Cortina-Borja, Copula models for record linkage
14.15 Stijn Vansteelandt, Estimating direct effects in cohort and case-control studies
14.45 Catherine Quantin, Linking multiple and heterogeneous databases: confidentiality and patient identification issues

15.15 Tea and biscuits + posters 

15.45 Ismail Ahmed/Pascale Tubert, Signal ranking-based comparison of automatic detection methods in pharmacovigilance: a hypothesis test approach
16.15 Daniela De Angelis, A Bayesian approach to estimating disease prevalence using information from multiple sources: HIV and HCV in England and Wales
16.45 end 

19.00 Drinks and dinner at the Mercure Parkside

Thursday 21st May 

09:00 Bendix Carstensen, Practical use of Lexis diagrams in the analysis and routine reporting from population registers
09.30 Paddy Farrington, Case series methods for the analysis of clinical and linked data
10.00 Hans van Houwelingen, Modelling long term survival data: beyond the proportional hazards model 

10.30 Coffee and pastries + posters 

11.00 Bianca De Stavola, Sensible uses of linked registry data when either outcome or exposure data are missing
11.30 Chris Jackson, Modelling combinations of population and survey data in epidemiological studies
12.00 Marie Reilly, Assessment of bias in familial risks estimated from population register data

12.30 Lunch + posters 

14.00 Paul Boyle, The value of linking lives through time for health research
14.45 Roundtable: Liam Smeeth, Using computerised clinical data for epidemiological research: problems and partial solutions; Carol Dezateux, Enhancing capacity: people, methods, infrastructure
15.15 Roundtable discussion
15.45 Conference closes

Report an error on this page