Ambulatory Care HER Update

1

• Health Information Technology Why Physicians
  Are Holding Back
  (and What Could Be
  Done to Change This)
• Biomedical Informatics Course NLM
• September 26, 2008
• Peter Basch, MD, FACP
• Medical Director, Ambulatory Clinical Systems

2
My perspective

• Practicing physician early adopter of HIT
• Using EMR in a small practice setting since 1997
• Using ePrescribing since 2002
• Medical Director Ambulatory Clinical Systems,
  MedStar Health
• Clinical leadership for MedStars EMR
  implementation
• Clinical leadership for MedStars HIE projects
• Chair, Maryland Task Force on EHRs
• Immediate Past Co-Chair, Physicians EHR
  Coalition
• Board Leadership Council member, eHealth
  Initiative
• Member, Medical Informatics Subcommittee, ACP

3
Overview

• Health IT a quick fix for suboptimal quality
  and uncontrollable costs
• My take on
• ePrescribing
• Clinical decision support
• A roadmap to integrated health IT adoption and use

4
Why bother?
5
And too expensive
6
And without sufficient value
7
HIT to the Rescue!
8
And not just the bottom line for health plans
9
If HIT can fix all that (and save
all that money for everyone)

• Easy to understand why politicians and policy
  makers are so interested
• More difficult to understand why the call for
  mandates has not come sooner and been louder
• And why physician adoption is relatively flat

10
HIT (per se) doesnt actually fix all that
11
Linder was not the first to study the correlation
between EMRs and quality improvement

• No correlation between EMR use and improved
  outcomes in diabetes
• (OConnor, PJ et al., Ann Fam Med, 2004)
• No improvement in lipid management after 5 years
  of using an advanced EMR
• (Siemienczuk, J et al, Am J of Managed Care,
  2004)
• Negative correlation between diabetes measures
  and EMR use
  (Crosson, JC, Ann Fam
  Med, June 2007)

12
But on the other hand

• Linder et al. used some measures of quality that
  many of us could have told him would be unlikely
  to be demonstrated with an EMR
• Recommended antibiotics (or their avoidance)
  based on diagnosis
• Avoiding certain medications in the elderly
• Avoidance of urinalysis during routine physicals
• Most current EMR research has studied EMR
  technology pre-2004
• Without almost any decision support
• Most EMR research has looked at use of EMRs in
  fee-for-service settings
• Without incentives to use whatever decision
  support was present

13
GAO Report on HIT

• GAO concluded that the underlying economic
  assumptions / calculations were suspect
• While they accepted the potential benefits of HIT
  as real, they felt that realizing those benefits
  required several other steps to occur which
  were far from certain

14
Two examples of questionable assumptions

• 11 of savings returned to physicians / HIT
  purchasers (89 of savings to payers, employers,
  patients) CITL report on ACPOE benefits
• Assumption 11 of MD revenue from global
  capitation where any savings achieved were
  dollar-for-dollar returned to the physician
• Global capitation is NOT 11 of revenue
• Dollar-for-dollar return is not a practice that
  exists anywhere in the US

• 87B/yr return for interoperability (where 40 of
  return is to providers) - CITL report on HIEI
  benefits
• Assumption projected benefits from not having to
  complete forms was additive (e.g., cost to
  complete a lab request was 10 for a CBC, 20 for
  a CBC and lipid panel, 30 for CBC, lipid panel,
  and ALT)
• To achieve those savings, an average MD would
  have to eliminate gt 100 of staff

15
GAO Report on HIT

• GAO concluded that the underlying economic
  assumptions / calculations were suspect
• While they accepted the potential benefits of
  HIT, they concluded that realizing those benefits
  required several other critical steps which
  were far from certain

16
What other steps necessary?

• Transformed healthcare system, including
• Defragmentation of healthcare delivery
• Aligned incentives for quality care

17
Making magic or magical thinking?
18
Whos to blame for this magical thinking?

• Convergence of
• Rapidly growing need to fix what is broken
• Increasing awareness of problems re quality and
  safety
• Increasing awareness of IT as a solution
• Over-promotion of HIT by early adopters and
  others
• Failure of people who should know better to think
  critically
• Potential for large new sources of profit in the
  business of healthcare

19
Removing the magical thinking quality
improvement

• HIT at best, presents an enabling infrastructure
• HIT could improve quality
• HIT could reduce certain errors
• HIT could reduce costs
• but only if
• It is sufficiently mature
• The healthcare system in which it is implemented
  is not dysfunctional / fragmented
• Physicians are incented to achieve the endpoints
  of quality, safety, and global efficiency
• HIT could just as easily
• Worsen quality
• Increase errors
• Increase costs

20
Removing the magical thinking ROI

• The EMR per se is ROI agnostic
• Modeled ROI with optimal use results in close
  to 100 of savings accruing to health plans
• No evidence that HIT would be optimally used
• Physician ROI may occur from
• Decreasing staff
• Decreasing / eliminating transcription
• Right-coding visits
• Where care is paid for differently (capitation /
  subscription, meaningful P4P, etc.) ROI may be
  less dependent on end-user implementation

21
Why physicians are holding back?

• HIT is expensive to purchase and maintain
• Difficult to implement and learn
• Has little intrinsic value without being used in
  the context of a transformed healthcare system

22
ePrescribing

• Was introduced (as a gadget / gimmick) in the
  midst of the dotcom bubble as a very inexpensive
  (free) approach to get doctors to adopt
  technology
• Disappeared for a while with the dotcom crash
• Re-appeared several years ago new and improved
  under the banner of medication safety

23
ePrescribing Benefits Legibility
24
Formulary management
25
Other ePrescribing benefits

• Legibility
• Formulary management
• Ease of renewals
• Medication history
• Fill history

26
Even though ePrescribing is cheap and easy
physicians are slow to adopt

• Assumptions are (again) incorrect / overstated
• Legibility is not the primary cause of medical
  errors not a significant cause of deaths from
  medication errors (meaning fixing illegible
  prescriptions doesnt actually do very much)
• And the fix for illegibility may be worse than
  the problem
• Concern that drop-down errors from ePrescribing
  might be more deadly than illegible prescriptions
• Medication history / fill history not as useful
  as portrayed
• Standalone ePrescribing may increase errors
• Fosters parallel workflow / shadow records
• ? which is the correct medication list
• No drug-condition, drug-lab decision support
• Insufficient context for many prescribing /
  renewal activities many (most) MDs think
  ePrescribing should only be done as part of an
  EMR implementation (which is why ePrescribing in
  vivo is neither cheap nor easy)

27
What about med history?

• Electronic medication history is typically a
  chronologic list of prescriptions filled and paid
  for by insurance
• Does not take into account
• Directions
• Medications discontinued
• Medications paid for with cash
• Medications actually taken
• Bottom linemedication history is incomplete /
  potentially misleading
• Should be thought of ONLY as a vehicle for an
  informed conversation
• Potentially helpful in ER situations where
  patient is unable to give a history

28
What about fill history?

• Fill history shows
• Prescriptions sent but not picked up by patient
• Prescriptions not refilled timely
• Raises questions of medication non-compliance
• What should / must a physician do with such
  reports?
• Discuss during visits?
• Call patient immediately?
• Have staff regularly track prescription use?
• And how will patients perceive this?
• Helpful / intrusive and creepy?

29
Standalone ePrescribing

• Physicians with no systems in place or
  e-readiness will be incented (and/or forced) to
  adopt ePrescribing, which for many will be
  standalone ePrescribing systems
• (Paper records combined with ePrescribing)

30
Even though ePrescribing is cheap and easy
physicians are slow to adopt

• Assumptions are (again) incorrect or overstated
• Legibility is not the primary cause of medical
  errors not a significant cause of deaths from
  medication errors
• Concern that drop-down errors from ePrescribing
  might be more deadly than illegible prescriptions
• Medication history / fill history not as useful
  as portrayed
• Standalone ePrescribing may increase errors
• Fosters parallel workflow / shadow records
• ? which is the correct medication list
• No drug-condition, drug-lab decision support
• Insufficient context for many prescribing /
  renewal activities many (most) MDs think
  ePrescribing should only be done as part of an
  EMR implementation (which is why ePrescribing in
  vivo is neither cheap nor easy)

31
Perhaps ePrescribing is the wrong target!
Uses HIT to enhance a paradigm shift in
medication management
Digitizes existential prescribing

• ePrescribing

• Knowledge-based medication management

• Legibility
• Drug-drug interaction
• Drug-allergy checking
• Drug-sex / age
• Formulary checking
• Interconnected to pharmacy system
• NO drug-condition checking
• Inclusion
• Exclusion
• NO drug-lab checking
• NO context surrounding medications and med
  history
• NO single medication list

• Legibility
• Drug-drug interaction
• Drug-allergy checking
• Drug-sex / age
• Formulary checking
• Interconnected to pharmacy system
• Drug-condition checking
• Inclusion
• Exclusion
• Drug-lab checking
• Rich context surrounding medications and med
  history
• Single medication list

• Most appropriate drug for the condition / disease
• Most cost-effective
• Right dose / formulation for that patient

32
Clinical Decision Support (external annoyances)
33
Clinical Decision Support (CDS)

• Alerts
• Drug-drug interactions
• Drug-allergy reactions
• Drug-sex alerts
• Drug-condition alerts
• Drug-age alerts
• Drug-lab alerts

• Prompts and Reminders
• Preventive care
• Chronic care
• Drug monitoring
• Calendared testing

34
CDS Passive
35
CDS Active
36
CDS granular / informative
37
CDS granular / actionable
38
My roadmap integrating CDS and ePrescribing
into Workflow
39
Passive CDS relevant, actionable, and polite
40
Granular / actionable CDS for diabetes
41
Granular / actionable CDS for smoking cessation
42
Granular / actionable CDS for prostate cancer
screening
43
Benefits of integrated CDS

• Makes it far more likely that patients will get
  better access to timely preventive and chronic
  care services
• Makes it far more likely that physicians will
  dramatically improve their quality scores with
  health plans, employers, consumer sites
• Presentation makes the prompts / reminders
  function more artfully than earlier iterations
• Presentation that allows for action and
  documentation makes protocol fulfillment take
  less time than non-integrated approach

44
Integrating ePrescribing into the EHR messages
from pharmacies
45
Simultaneous prescribing to a pharmacy and PBM
46
Benefits of integrated ePrescribing

• Satisfies most of the concerns re standalone
  ePrescribing
• Single (more likely to be accurate) med list
• Prescribing / renewing done in full context
• Does not present itself as outside of normal
  workflow
• Lends its to knowledge-based medication
  management

• Inbound messaging
• Clear savings of staff time
• May in itself pay for EHR maintenance / support
• Outbound prescribing
• Convenience for patients
• Simultaneous prescribing to pharmacy and PBM
• Works in progress
• Medication history
• Fill history

47
Summary problems

• Health IT huge potential, but oversold to
  policy makers and politicians
• Lack of rigorous analysis does not just lead to
  overselling it interferes with actual problem
  solving (and helping health IT to mature)
• Early adopters bought the vision / potential of
  health IT everyone else is buying based on
  realistic assessments

48
Summary solutions

• Stop magical thinking / stop focusing on the
  health IT as more than infrastructure
• Focus on the goals of transformed health care
  delivery reformed reimbursement
• Physicians and other providers will then see
  health IT as a wise investment
• Adopt appropriately
• Use optimally
• Push for enhancements to health IT that best
  serve all stakeholders