Ehealth

1
E-health New benefits, new risks Healthconnect
Hobart 2007
Enrico Coiera e.coiera_at_unsw.edu.au www.chi.unsw.e
du.au
2

• George Pompidou
• President of France 1969-1974
• There are three roads to ruin women, gambling
  and technicians.
• The most pleasant is with women, the quickest is
  with gambling
• but the surest is with technicians.
• (with thanks to Michael Kidd)

3
Its not just about user consultation

• Its about active participation
• Sometimes
• Clinicians are interested observers.
• Clinicians are stakeholders - whenever the issue
  will impact in our sphere e.g. EHR standards,
  consent legislation
• Clinicians are 1st class partners - e.g. whenever
  we are dealing with working systems, clinicians
  must help design the work. Forget them and you
  will fail.
• Clinicians are the leaders - whenever the problem
  is theirs
• We can make the same argument for the inclusion
  of consumers, private healthcare and industry etc

4
The price for not participating

• Many large scale IT implementations struggle or
  fail e.g.
• Kaiser Permanente wrote off US0.5 billion in
  clinical IT systems that didnt fit workflow
• In 1989 NSW Health spent gt 110 million on
  hospital IT, withdrawn by 1995 because took
  clinicians longer to use (JAMIA,112-124, 1997).
• Why? Its not about the IT. Its about the whole
  system. We are delivering complex integrated
  health services, not IT solutions.
• Sometimes IT, underestimating the complexity of
  health, is not a good team player.

5
Where should clinicians lead?

• Evidence-based practice
• Safety and quality of care
• Both need clinician and consumer leadership, and
  ultimately will depend heavily on e-health to
  provide solutions

6
Evidence-based practice

• A new article is added to medical literature
  every 26 seconds.
• Clinicians knowledge decays with years since
  graduation (Evans et al., 1984)
• 2/3 of 8.5 p.a. growth in health costs driven by
  demand for new technologies but only 21
  supported by evidence of benefit
• 17 hospital admissions result in adverse event,
  5 of which result in death 14k p.a often due
  to poor information

7
(No Transcript)
8
(No Transcript)
9
(No Transcript)
10
Clinical Information Access Program (CIAP)

• Introduced in 1997
• 24 hour, online evidence retrieval system at the
  point-of-care
• www.ciap.health.nsw.gov.au
• Available to approx. 55,000 clinicians in NSW

11
Day and time of evidence use
12
Percentage of admissions and evidence searches by
day
13
Monthly rate of single source database use by
public hospitals in NSWRate per 100 clinicians
14
(No Transcript)
15
Controlled Laboratory Trials

• 75 clinicians - 26 hospital doctors, 18 GPs, 31
  clinical nurse consultants)
• Answer 8 medical problems
• Decision accuracy - 21 improvement
• Pre-search 29 correct
• Post-search 50 correct
• Time to correct answer - 51 improvement
• QC 4.5 min
• No profiles 6.8 min

16
Results
17
Evidence-based practice

• What should clinicians be working on?
• Training working clinicians in the use of on-line
  resources.
• Sophisticated technologies require sophisticated
  users
• Google considered harmful to your health
• Culture change comes from within
• Finding a sustainable national model for
  collecting and disseminating evidence to ALL
  specialties and consumers - a national e Health
  Library

18
(No Transcript)
19
Should clinical software be regulated?
20
How safe and effective are our systems?

• Inventory of 36 Australian EDSS for Task Force
  (2002)
• Represented approx 230 million
• Two thirds with public funding

21
Evaluation methodologies
Qualitative
Case study
Before/after sample
RCT
Not done
22
Process measures used
Improved
No change
Not measured
23
Evaluating e-prescribing safety in general
practice

• 4 commercial UK prescribing systems evaluated
  against 18 significant scenarios
• N appropriate alerts 4,7,4,3 (max 18)
• The safety features of computing systems
  currently in use in about three quarters of UK
  general practices have clinically important
  deficiencies. All may fail to warn when a
  warning is expected, potentially creating a
  health hazard to patients.
• BMJ  20043281171-1172 (15 May)

24
NPS Drug interaction prompts study

• 4 GP software packages evaluated
• Drug interactions tested
• Newer agents with significant interactions
• Recently reported in ADRAC bulletin and well
  documented in the literature
• Drug interactions in the elderly

25
Expected Vs actual prompts
26
Automation biases

• Errors of omission - events missed because the
  not drawn to attention of the user
• 59 accuracy on the omission error events,
  compared to 97 for non-computer users (Skitka,
  1999)
• Errors of commission - users do what DSS tells
  them, even if contradicts training / data.
• Performance with aid even worse on the commission
  error opportunities, with an accuracy of only 35
  (Skitka, 1999)
• Errors of dismissal - DSS advice ignored
• eg physicians overrode 89 high severity drug
  interaction CPOE alerts (Weinggert et al. 2003)

27
Explaining automation bias

• Devolving responsibility to computer
• Out of loop unfamiliarity, arising from less
  attention to raw data, resulting in poor
  situational awareness
• Fixes for automation bias
• Short term impact only from training
• Accountability for performance seems to lead to
  greater attempt to verify decision aid
  recommendation

28
Four Stages in the evolution of system design
29
Interruptions and e-PS

• 570 US hospitals, 2003
• 20 medication errors involved computerisation /
  automation
• Computer data entry errors the 4th leading cause
  of medication error
• 60 of these due to user distraction
• (US Pharmacopiea, 2005)

30
Safety and quality of care

• What should we be working on?
• Lobby hard for real to support complex systems
  research in health services - shift focus from
  the cell to the system.
• Training working clinicians in the use of DSS,
  especially e-prescribing.
• Understand the benefits and risks
• Understand cognitive biases that lead to error
• Understand how the machine works and breaks
• Lobby hard, and work with industry, for national
  mechanisms to ensure safe e-prescribing systems

31
National EDSS Evaluation Guidelines

• Overall Aims
• To promote evaluation of EDSS
• To establish a common evaluation framework to
  permit comparison and sharing of experiences
• To provide guidance on evaluation methods for
  common questions
• http//www2.chi.unsw.edu.au/edsse/wrapper.php

32

• Thank You
• e.coiera_at_unsw.edu.au