NESSTAR and Eastern Public Health Observatory (erpho) Julian Flowers.
Post on 13-Jan-2016
216 Views
Preview:
Transcript
NESSTAR and Eastern Public Health Observatory (erpho)
Julian Flowers
Outline
• What PHOs do
• Where does NESSTAR fit in?
• Implementing NESSTAR in PHOs – issues
• Where next…?
PHOs
• Regional bodies – NHS/DH funded• Serve public health community by improving
access to data/ information/ expertise• Cataloguing data sources => metadata• Informatics = languages/ controlled vocabulary/
taxonomy etc…• Interoperability = eGMS/ standards• Data production/ secondary analysis• Reports• Education and training
Who are our users?
• Primarily– Ourselves– Information analysts in PCTs/ SHAs etc– Directors of Public Health and their teams
• Secondarily– Other health staff– Non-NHS– Academics– DH– The public….
erpho and data
• Trying to improve access to routinely available data
• Initially developed website as a catalogue of datasets i.e. database of metadata
• => signposting function• Access to microdata from a variety of sources
=>Increasingly asked to provide data and analysis• Need tools to improve both access to and
accessibility of information
NESSTAR – why we chose it
• Ability to publish micro-data AND cubes
• Ability to handle large datasets
• User friendly interface
• Metadata
• Searching within the server e.g. for variables – makes analyses much easier
=> Improves both access and accessibility of datasets
Implementation
• Customisation• Bespoke version of server• Linked in to PHORMS – PHO website for
– Resource discovery– Studies and variables published on server searchable through
the website – Access control
• Uses an authentication service for “one-stop” password and user ID control
• Used PHITS catalogue• Data preparation• Training
What is published where
• Use specialist catalogue for health data
• Publish data from a variety of sources– Republish routinely available data from
national bodies– PHOs have access to record level data
• Publish analyses of hospital activity, mortality etc.• Also anonymised extracts of datasets for selected
users
Issues
• Disclosure and disclosure management• Access control = needs to vary by user and by
dataset– Data in public domain should be unrestricted– Data not in public domain but publishable should be
unrestricted (FOI)– Data not in public domain but not routinely
publishable should have restricted access, disclosure control…
• Access control = how?
Issues (2)
• Catalogue flexibility• Searching through the web vs. searching on the
server• URLs• Speed• Complexity of some of the variables• Data preparation• Cubes versus “studies” – can do different things
with each and may need to publish in both forms…
Where next?
• Revamping links between server and website
• Revamp health catalogue – needs to be more flexible
• Retaining existing access control• ? Mapping• More user friendly interaction• Careful consideration of what we publish
in what form to whom
top related