System Watch: A webbased system to monitor and predict pressure in the Scottish health service

1
System Watch A web-based system to monitor and
predict pressure in the Scottish health service

• Helen Brown
• University of Edinburgh
• Information and Statistics Division (NHSScotland)

2
Aims

• Monitor and predict pressure in the NHS Scotland
  in real time
• Identify vulnerable groups of patients and inform
  clinicians
• Provide a patient information system

3
Indicators of pressure

• Emergency admissions
• Beds occupied by emergency admissions
• Calls to NHS24 (relating to infectious illness)
• Flu spotter rates
• Ambulance emergency call-outs
• Local monitoring information
• (Weather, air pollution)

4
Real-time processing

• Data feeds-
• Individual in-patient records from all Scottish
  hospitals and NHS24
• Flu surveillance rates and 999 calls frequencies
• Sent daily/weekly
• Data processing, calculation of predictions,
  production of graphs and tables as web pages
• Pass information to System Watch website

5
Predictions and trends

• Hospital data modelled to construct-
• Seasonal trend Based on seasonality, day of
  week, public holidays.
• Real time prediction As seasonal trend,
  additionally using current hospital emergency
  activity and flu rates.

6
SHOW WEBSITE
7
Identifying vulnerable groups of patients

• Linkage of patient records allows their histories
  to be tracked, selected groups of patients can be
  identified
• For example, repeated emergency admissions in
  elderly patients can be identified as they arise
• Hospital staff and/or GPs notified by email
• Potential to identify other groups, eg
• repeated callers to NHS24
• repeated admissions for self-harm, or in children

8
Providing patient-based information

• Tool to provide in-patient histories and NHS24
  call details on request
• Tool to inform designated person via email/mobile
  when a patient is admitted to hospital
• Confidentiality issues to be addressed before
  tools are implemented

9
Planned development

• More data
• More frequent updates (eg daily), eventually
  automate data feeds
• Improve accuracy of predictions (eg include NHS24
  information consider more complex models)
• Real-time linkage of records
• New products utilising real-time data

10
More real-time data

• NHS data AE, deaths, GP records (GPASS),
  prescriptions, disease registers, maternity,
  children, test results
• Non-NHS health data
• private health care
• data from patient monitoring systems (eg asthma
  project)
• hits and information entered onto personalised
  health related websites (advice sites, managed
  clinical network sites)
• Other person-based data social services, police,
  schools
• Environmental weather, air pollution

11
Real-time linkage

• Linkage of records on same person
• 3 approaches
• Record unique patient identifier (CHI in
  Scotland)
• Link pairs of datasets using non-unique
  identifiers eg name, dob, postcode (and CHI when
  recorded) - probability matching
• Link CHI to all datasets

12
New products using real-time data

• Integrated care records
• Interactive interface to provide tailored
  information
• Workflow management tools, eg scheduling
  operations
• Patient monitoring and notification tools
• Patient level prediction tools, eg children at
  risk, risk of emergency admission

13
Accessing System Watch

• Website www.show.scot.nhs.uk/systemwatch
• Login system
• Password watch
• (Access to NHSNet required)