Optimising the use of routine administrative health records: the role of data science methods

21 May 2019, 10.00 AM - 21 May 2019, 2.00 PM

Hepple Lecture Theatre, School of Geographical Sciences, University of Bristol, BS8 1RL

Workshop aim

The aim of the workshop is to bring together researchers from Population Health Sciences and data scientists/mathematicians/engineers from across the University of Bristol to form new collaborations to begin to answer complex questions using routine administrative health records. Health care is increasingly complex, and analytical techniques that utilise more of the available information in routine data sources could begin to answer more complex questions, thus optimising the delivery of safer, efficient and more effective care.

What to expect

The morning will include presentations on accessible and available databases for research, and data science methodologies, including free-text analysis, high dimensional data structural analysis, and machine learning, from experts at the University of Bristol.

Researchers from Population Health Sciences will then present research challenges - clinical, health service research or public health questions that use available routine health date that require novel (for epidemiology/health service research) methodologies such as neural networking, free text analysis, data mining, inductive logic programming and decision support and recommender systems. Challenges are broad, and are an opportunity to develop new methodologies that could then be applied elsewhere.

The presentations will be followed by round table discussions and a working lunch, for ideas to be discussed and developed in an informal environment.

The morning is open to everyone, including students who are looking for PhD projects, and early career scientists. We hope that by bringing researchers together, we can share our expertise and identify new and exciting opportunities for cross-discipline collaborations.

Structure of the day

Welcome  10.00-10.05  Introduction/aims of the morning  Rupert Payne, Jonathan Lawry 
Session 1  10.05-10.35 

Health data
A clinical perspective on the utility of health data.
Examples of national datasets used in Bristol (CPRD, HES).
Examples of local dataset used in Bristol (ALSPAC)


Rupert Payne, Rachel Denholm, TBC



Free text mining and data analysis
Challenges in High Dimensional Data Structural Analysis
Machine Learning

James Freeman, Song Liu,

Raul Santos Rodriguez 

Coffee  11.10-11.30      
Session 2  11.30-12.10 

Open Challenge Questions:

3-min presentations on 10 health research ‘challenges’

Various PHS staff 
Session 3  12.10-13.00 Round table discussions  ALL 
Lunch 13.00-13.50 Including further networking and discussion  

Closing summary

13.50-14.00 Closing summary and future steps  Rupert Payne, Jonathan Lawry 

Data Week 2019

This event is part of Data Week 2019, run by the Jean Golding Institute

Any questions? Please contact jgi-coordinator@bristol.ac.uk


Please register for this event via Eventbrite 

Optimising the use of routine administrative health records: the role of data science methods

Edit this page