The impact of information technology on the quality and safety of health care Systematic literature

1
The impact of information technology on the
quality and safety of health care Systematic
literature overview and synthesis
Professor Azeem Majeed Head of Department
Professor of Primary CareDepartment of Primary
Care Social Medicine Imperial College London
2
(No Transcript)
3
Why e-health?

• Increasing burden of care (e.g. increased life
  expectancy, long term conditions, range of
  treatment options)
• Rising healthcare costs
• Continuing health systems inefficiencies
• Variations in quality of care
• High prevalence of medical errors
• Greater public scrutiny of government spending
• Patients and the public want a greater say in
  decisions about their health healthcare

4
The emergence of e-health

• Rapid IT developments across all sectors, but
  diffusion still relatively limited in health care
  sector
• Need to capitalise on potential of IT to increase
  efficiency, effectiveness and safety
• But implementation of IT also has potential to
  introduce new risks errors associated with
  technology or users
• New IT solutions often rapidly introduced, on a
  large scale, despite limited evidence of
  effectiveness or safety
• Many IT innovations remain un(der)studied and
  evidence of their effectiveness is inconsistent

5
Scope of the review

• Commissioning brief emphasised areas of high
  relevance for CfH, namely, hence areas targeted
  in this review
• Storage and retrieval of medical records and data
• Health information exchange and interoperability
• Electronic health records
• Computer history taking systems
• Supporting professional decision making
• Computerised decision support systems
• ePrescribing
• Socio-technical considerations for eHealth
  development and deployment
• Human factors
• Effective implementation and adoption of eHealth
  applications

6
Conceptual work

• Purpose
• Develop theoretical grounding by profiling the
  concepts of eHealth, quality and safety.
• Analytic framework draws on
• Core applications of eHealth, represented within
  the conceptual model developed by Pagliari et al.
  
• Taxonomy of patient safety risks developed by
  WHO/JCAHO Key elements of healthcare quality
  frameworks
• Headline deliverables of NHS Connecting for
  Health and other key elements of the programme

7
Mapping healthcare quality

• Multiple conceptual approaches (e.g. Campbell,
  Donabedian, Shekelle, Schuster)
• Key attributes of quality
• Effectiveness (outcomes of care,
  appropriateness/evidence-basis of care given)
• Efficiency (streamlining care processes, reducing
  costs)
• Acceptability (satisfaction, patient-involvement)
• Equity (reducing variations in practice)
• Accessibility (speed and ease of obtaining care
  or advice)

8
Quality map
(Derived from multiple sources, e.g. Campbell,
Donabedian, Shekelle, Schuster)
9
(No Transcript)
10
(b)
CAUSE
SYSTEMS
HUMAN (Error)
Technical
Organisational
Patient
Practitioner
External factors
Management e.g. staffing/training, safety
financing
IT planning procurement
Facilities e.g. design, malfunction obsolescence
Skills based (execution)
Data entry, software use
Culture e.g. around information sharing or event
reporting
Towards change innovation
Patient factors
Rule-based (retrieval)
IT usability, accessibility, dependability, QI
loops
Data extraction
e-protocols (e.g. referral), coding QoF, IT
standards
Negligence Recklessness Intentional rule
violations
Protocols/Processes e.g. documentation,
incentives, standards
IT skills, ability to recognise good poor
quality info. Use of unsecure e-comms for
sensitive info.
Knowledge-based (interpretation)
Knowledge Transfer e.g. training supervision
around safety
IT skills. Process redesign
Lack of tools/procedures for tracking, auditing,
alerting (e.g. Shipman)
Data interpretation (e.g. images, reports,
records) information, evidence, decision support
11
(c
Due to faulty infor-mation, interventions or
confidentiality breaches
Morbidity and mortality due to poor CDSS, e-comms
or EHR integrity
Confidentiality negligence suits
Poor adoption of technology
Cost to modify or replace IT re-train
staff Legal settlements. Lost bed days waiting
times
12
Challenges to synthesising the evidence

• Large and rapidly expanding literature, poorly
  indexed, of variable quality
• Hence we used the following approach
• Clarifying definitions, description and scope for
  deployment to reflect on potential risks
  benefits
• Identifying empirically demonstrated benefits and
  risks, using exemplar subject areas and/or
  detailed case studies with relevance to CfH
• Highlighting clinical, policy and research
  implications

13
Example Electronic Health Records

• Range from simple digital databases to complex
  systems integrating electronic communications
  (e.g. ePrescribing) and active knowledge support
  (e.g. clinical decision support).
• Theoretical Benefits

14
Electronic Health Records ctd

• Theoretical Risks
• Vulnerability to power cuts and system failures
• Increased time for some data entry
• Provider resistance to change
• Lack of interoperability
• Security/privacy risks
• Poor data quality (coding)
• Reliance on possibly incomplete patient data
• Increased resource utilisation at early stages

15
Electronic Health Records ctd

• Empirical evidence of benefits
• Good evidence of
• Improved legibility
• time saving for some professionals (nurses)
• facilitation of audit, performance management
  secondary analysis of routine data
• Moderate evidence of
• improvements in clinical decision making for
  preventive care (complicated by integration of
  CDSS)
• Little evidence of impacts on
• clinical outcomes
• Cost (some efficiency gains in ePrescribing)
• Patient safety
• Empirical evidence of risks
• Increased doctor time for data entry and
  retreival

16
Example Supporting professional decision making

• Clinical decision support
• and
• e-supported prescribing

17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
Example Electronic Prescribing

• Theoretical benefits
• Increased legibility
• Standardisation of prescribing
• Patient specific support for prescribing
• Reduced lost orders
• Instantaneous transmission to pharmacy
• Instant reporting that item is out of stock
• Alerts for unfilled and/or non-renewed
  prescriptions
• Improved documentation
• Tailored prescriber feedback
• Facilitated clinical governance activities
• Cost-savings

24
Electronic Prescribing

• Empirical evidence of benefit
• ? Obliteration of illegibility
• ? Reduction of prescribing errors in inpatients
• ? Increased time for patient care (pharmacists)
• ? Reduction of turn-around times
• ? Determining more appropriate drug dosage
• ? Increased corollary orders
• ? Reduction of preventable adverse drug events
  (ADEs) in inpatients
• ? Reduction of prescribing errors in outpatients
• ? Guideline adherence by outpatient providers
• ? Cost-benefit ratio

25
Electronic Prescribing

• Empirical evidence of risk
• ? Poor specificity of alerts (excessive alerts)
• ? Poor sensitivity of alerts (missed alerts)
• ? Alert fatigue (non-adherence)
• ? Increased work load
• ? Poor system fit
• ? Introduction of errors (human and machine)

26
Overarching observations I

• Ripe environment for heathcare IT
• Potential to ameliorate burden of care caused by
  changing health demographics, proliferation of
  treatments costs
• Rising public expectations for service quality
  and efficiency
• Rapid advances in IT
• European lead in eHealth industry
• Gap between theoretical demonstrated benefits
• While seminal reports recognise the potential
  impact of eHealth on quality and safety, numerous
  human and technical barriers need to be overcome
  before these systems are embedded sufficiently
  for this potential to be realised. Most eHealth
  technologies are supported only with face
  validity or weak empirical evidence. There is a
  need for more evaluation across the lifespan of
  technology design and implementation.

27
Overarching observations II

• Variable quality of evidence limits
  interpretation
• Relatively few eHealth trials have evaluated
  safety outcomes
• Lack of primary research assessing quality
  outcomes (implicit assumption that benefits are
  obvious)
• Studies demonstrating greatest benefits often
  come from academic medical centres and concern
  home grown systems, so results may not be
  generalisable to less conducive environments.
  Joint roles as developer/evaluator/user may
  introduce bias.
• Inconsistent use of outcome measures make it hard
  to generalise across studies
• Poorly theorised interventions and studies
  (failure to anticipate complexity)
• Poor outcome definition and measurement
• Overestimating likely effect sizes.
• Inappropriately short evaluation timescales (not
  enough time for interventions to bed-in and
  improvements demonstrated)

28
Overarching observations III

• Vast and expanding body of literature is poorly
  indexed, appraised and ordered
• Large body of work at intersection between
  eHealth, quality and safety
• Overlap between different eHealth applications
  variation in contexts of implementation make it
  difficult to produce meaningful taxonomic
  frameworks and assess likely effectiveness
  generalisability
• Inadequate attention to human and socio-cultural
  factors
• Human factors (e.g. poor usability) may
  compromise the fitness-for-purpose of otherwise
  sound applications e.g. evidence that many
  doctors routinely ignore e-alerts where their
  value is not recognised. Results obtained in
  controlled trials may therefore not generalise to
  real world settings. Greater attention to design
  considerations, such as tailoring of support and
  preventing over-rides for critical alerts could
  improve this situation.
• Contextual factors (local culture,
  rewards/incentives and previous technical
  history) can strongly influence the transfer
  potential of technologies and studies should
  document these softer influences to aid
  interpretation of results and inform future
  implementations.

29
Conclusions

• Greater attention to sociotechnical factors in
  the design, implementation and evaluation of
  eHealth interventions is necessary to maximise
  their potential
• Demonstrating their potential value for improving
  quality, safety and efficiency is dependent on
  the application and transparent reporting of
  methodologically sound and theoretically robust
  evaluation research
• Such evidence is of high value to suppliers and
  policymakers (e.g. for business case
  demonstration) as well as health service
  providers and patients.

30
Some key issues for research

• Giving patients online access to their medical
  records. What are the confidentiality, technical,
  ethical and legal issues? How will it impact on
  the quality of care and accuracy of electronic
  medical records?
• Giving the public the information they need to
  make informed choices about their health and
  their healthcare. Use of web technologies, social
  networking etc.
• Self management programmes and patient
  empowerment for people with long term illnesses.
• Use of information technology to improve quality
  of care, for example through the use of
  computerised decision support systems.
• Developing capacity to process very large volumes
  of data for public health surveillance quickly
  and efficiently to allow early detection of
  threats (e.g. flu outbreaks, adverse drug
  reactions).

31
Acknowledgements

• Thanks to the members of the team
• Dr Josip Car, Imperial College
• Ms Ashly Black, Imperial College
• Professor Aziz Sheikh, Edinburgh
• Dr Claudia Pagliari, Edinburgh