Evaluating the Massachusetts eRx Collaborative: Does It Improve Patient Safety and Reduce Costs

1
Evaluating the Massachusetts eRx Collaborative
Does It Improve Patient Safety and Reduce Costs?
Joel S. Weissman, Ph.D. MGH/Harvard Institute for
Health Policy

• AHRQ Conference Patient Safety and Health
  Information Technology Making the Health Care
  System Safer through Implementation and
  Innovation
• June 9, 2005

2
Context

• 8.1 million households have experienced a
  potentially dangerous medical or
  prescription-drug error (Commonwealth Fund)
• Crossing the Quality Chasm focused on IT as a
  tool to minimize human errors.
• The Medicare Prescription Drug Improvement and
  Modernization Act of 2003 (MMA) requires plans to
  be able to use e-Rx
• The emergence of low-cost Web technologies and
  the widespread use of palm-sized devices has made
  the move to e-prescribing a reality.

3
What I will Cover

• Overview of the federal grant
• The e-Rx Collaborative
• Tool Functionality related to Patient Safety and
  Cost Savings
• Early data on adoption
• Study Methods
• Next Steps

4
A Natural Experiment
In October 2003, Tufts Health Plan (Tufts HP),
Blue Cross Blue Shield of Massachusetts (BCBSMA)
and Zix Corporation collaboratively launched a
voluntary e-prescribing program in Massachusetts.
The program uses the PocketScripts device
provided by Zix Corporation.
5
AHRQ Grant E-Prescribing Impact on Patient
Safety, Use, and Cost

• Federal grant supported by the Agency for
  Healthcare Research and Quality (AHRQ)
• Partnership with Plans
• The project has two major aims
• Determine relationship between eRx and safety
• Determine relationship between eRx and cost,
  utilization, and outcomes
• Analysis using claims data

6
Study Personnel

• MGH
• Christine Vogeli, Ph.D.
• Tim Ferris, MD
• Rainu Kaushal, MD
• BWH
• Mike Fischer, MD
• Barry Blumenfeld, MD

• BCBSMA
• Megan Bell
• Jessica Fefferman
• Robert Mandel, MD

• Zix Corps (data provision, only)
• Danny Sands, MD

• Tufts Health Plan
• Jim Courtemanche
• Julie Newton

7
Study Questions

• 1) Do physicians who use e-Rx experience a lower
  rate of apparent poor prescribing habits compared
  with non-users?
• 2) Do physicians who use e-Rx increase use of
  generics, comply with formulary, and decrease the
  cost per prescription?
• 3) Do patients with selected chronic illnesses on
  e-Rx have fewer ED visits and hospitalizations?

8
The Proposal

• A collaborative effort to promote the use of eRx
  technology through widespread introduction of
  point of care technology in physician offices

The Players

• Outgrowth of BCBSMA and Tufts Health Plan
  individual ePrescribing pilot programs
• Neighborhood Health Plan joined in August 2004
• ZixCorp was initial technology partner, added
  DrFirst in 2005

The Goals

• Enhance patient safety
• Improve office efficiencies
• Increase provider and member satisfaction
• Improve formulary compliance
• Reduce pharmacy cost trend

9
ePrescribing Benefits

• Prescriber
• (Clinical Perspective)
• Mobile Rx writing
• Access to patients Rx history
• Drug-drug interaction alerts
• Allergy alerts
• Drug reference guide
• Reporting capabilities

• Prescriber
• (Administrative Perspective)
• Quicker refill process
• Reduced pharmacy calls
• Access to plan specific formulary
• Access to prior authorization forms
• Personalized frequently used Rx lists
• Electronic pharmacy address book

• Patient
• Access to patients Rx history
• Increased safety
• Reduced pharmacy wait times
• Increased level of confidence
• Improved medication compliance
• Predictable co-pay

• Health Plan
• Control increasing healthcare cost trends
• Improved formulary adherence

10
Program Design Free is not good enough

• Funding is key
• ePrescribing costs for an individual prescriber
  are estimated anywhere between 1,600-10,000 for
  the first year and 950-2,800 for subsequent
  years
• eRx Collaborative believes first year funding is
  critical to adoption
• Program subsidizes
• Handheld device
• 1 year ePrescribing application license
• Installation, Training and Support
• 6 months of internet connectivity where
  applicable

11
Functionality Drug-Drug Interaction (DDI) Alerts
12
Functionality - Checking the Formulary
13
Functionality - Dispensed Drug History
14
eRx Collaborative Adoption

• As of Q1 2005
• Over 2600 successful Program recruitments
• Over 2000 prescribers with the technology
  incorporated into their office
• 900 Active Prescribers
• 160,000 eRx per month

Active Prescribers Prescribers who averaged 5
or more scripts a week
15
Total Scripts and Total Active Prescribers (June
2005)
16
eRx Deployment Momentum
17
Practice Size (January snapshot)
18
Study Methods Data sources

• 2 years of claims data from THP and BCBSMA (April
  2003-March 2005) including
• enrollment
• inpatient and ambulatory claims
• dispensed drug claims
• 13 months (March 2004-March 2005) of data from
  Zix Corporation on prescriptions written using
  the PocketScripts Tool.

19
Basic Pre Post Design with Concurrent Control

• Pre- Post period based on subjective cut-off
  date, but will perform sensitivity analyses (and
  multivariate time series)
• A, B, C, D represent errors rates
• Is A gt B (i.e., do error rates decrease among
  adopters?)
• Is A-B gt C-D (i.e., is the pre-post reduction in
  error rates greater among adopters than among
  non-adopters?)

20
Dependent Variables Potential Errors

• Potential Drug-Drug Interactions (DDIs)
• potential, since we would not know if one
  actually occurred
• pair-wise analysis of concurrent medications
  using a module supplied by First Databank
• Apparent Dosage Errors and Duplicate Therapy
• range-checking module from First Databank
• outpatient alerts developed by Barry Blumenfeld,
  MD, and others at Partners HealthCare System
• Potentially inappropriate drugs for classes of
  patients

21
Dependent Variables Utilization and Outcomes

• Formulary compliance and use of generics
• Clinical endpoints
• e.g., reduced ED visits and hospitalization among
  patients with congestive heart failure, asthma,
  and diabetes

22
Analysis

• Units of analyses will vary, depending on
  question of interest
• the prescription
• the patient
• the physician
• Pre-post t-tests or paired t-tests
• Multivariate analysis
• control for physician propensity to adopt, and
  patient-level health status, demographic factors,
  and benefit differences.
• Interrupted time series (using repeated measures)

23
Results and Conclusions

• Forthcoming

"It is an art of no little importance to
administer medicines properly but, it is an art
of much greater and more difficult acquisition to
know when to suspend or altogether to omit
them" - Philippe Pinel 1745-1826
French physician and psychiatrist, considered
one of the founders of psychiatry physician to
Napoleon
24
End of presentations
25
eRx Collaborative Adoption

• As of Q1 2005
• Over 2600 successful Program recruitments
• Over 2000 prescribers with the technology
  incorporated into their office
• Over 40,000 electronic scripts sent during the
  last weekly reporting period in March 2005 a 41
  increase from the highest weekly script count of
  the previous quarter

Active Prescribers Prescribers who averaged 5
or more scripts a week
26
The Proposal

• Develop a collaborative effort to promote the use
  of eRx technology through widespread introduction
  of point of care technology in physician offices

The Players

• Outgrowth of BCBSMA and Tufts Health Plan
  individual ePrescribing pilot programs
• Neighborhood Health Plan joined in August 2004
• Initially partnered with ZixCorp as the
  technology partner, added DrFirst in 2005

The Goals

• Overall
• Enhance patient safety
• Improve office efficiencies
• Increase provider and member satisfaction
• Improve formulary compliance
• Reduce pharmacy cost trend

• 2005 Focus
• Provide multiple technology options
• Increase access to more formularies
• Program Evaluation
• Increase ePrescribing technology adoption and
  utilization!

27
Defining errors

• 1. Unauthorized drug
• 2. Extra dose
• 3. Wrong dose
• 4. Omission
• 5. Wrong route (e.g., right eye instead of left
  eye)
• 6. Wrong form (e.g., suppository instead of a
  tablet)
• 7. Wrong technique (e.g., injecting insulin into
  an incorrect site)
• 8. Wrong time (measured if administration
  occurred at least one hour before or one hour
  after scheduled time of dosage).

28
Test for interactions
29
Recruiting Momentum